Teradata Database SQL Reference Fundamentals Release V2R6.2 B035-1141-096A September 2006

ou juste avant la balise de fermeture -->

Voir également d'autres Guide Teradata : Teradata Database SQL Reference Fundamentals Release V2R6.2 B035-1141-096A September 2006The product described in this book is a licensed product of Teradata, a division of NCR Corporation. NCR, Teradata and BYNET are registered trademarks of NCR Corporation. Adaptec and SCSISelect are registered trademarks of Adaptec, Inc. EMC, PowerPath, SRDF, and Symmetrix are registered trademarks of EMC Corporation. Engenio is a trademark of Engenio Information Technologies, Inc. Ethernet is a trademark of Xerox Corporation. GoldenGate is a trademark of GoldenGate Software, Inc. Hewlett-Packard and HP are registered trademarks of Hewlett-Packard Company. IBM, CICS, DB2, MVS, RACF, OS/390, Tivoli, and VM are registered trademarks of International Business Machines Corporation. Intel, Pentium, and XEON are registered trademarks of Intel Corporation. KBMS is a registered trademark of Trinzic Corporation. Linux is a registered trademark of Linus Torvalds. LSI, SYM, and SYMplicity are registered trademarks of LSI Logic Corporation. Active Directory, Microsoft, Windows, Windows Server, and Windows NT are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries. Novell is a registered trademark of Novell, Inc., in the United States and other countries. SUSE is a trademark of SUSE LINUX Products GmbH, a Novell business. QLogic and SANbox are registered trademarks of QLogic Corporation. SAS and SAS/C are registered trademark of SAS Institute Inc. Sun Microsystems, Sun Java, Solaris, SPARC, and Sun are trademarks or registered trademarks of Sun Microsystems, Inc. in the U.S. or other countries. Unicode is a registered trademark of Unicode, Inc. UNIX is a registered trademark of The Open Group in the US and other countries. NetVault is a trademark and BakBone is a registered trademark of BakBone Software, Inc. NetBackup and VERITAS are trademarks of VERITAS Software Corporation. Other product and company names mentioned herein may be the trademarks of their respective owners. THE INFORMATION CONTAINED IN THIS DOCUMENT IS PROVIDED ON AN “AS-IS” BASIS, WITHOUT WARRANTY OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING THE IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NONINFRINGEMENT. SOME JURISDICTIONS DO NOT ALLOW THE EXCLUSION OF IMPLIED WARRANTIES, SO THE ABOVE EXCLUSION MAY NOT APPLY TO YOU. IN NO EVENT WILL NCR CORPORATION (NCR) BE LIABLE FOR ANY INDIRECT, DIRECT, SPECIAL, INCIDENTAL OR CONSEQUENTIAL DAMAGES, INCLUDING LOST PROFITS OR LOST SAVINGS, EVEN IF EXPRESSLY ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. The information contained in this document may contain references or cross references to features, functions, products, or services that are not announced or available in your country. Such references do not imply that NCR intends to announce such features, functions, products, or services in your country. Please consult your local NCR representative for those features, functions, products, or services available in your country. Information contained in this document may contain technical inaccuracies or typographical errors. Information may be changed or updated without notice. NCR may also make improvements or changes in the products or services described in this information at any time without notice. To maintain the quality of our products and services, we would like your comments on the accuracy, clarity, organization, and value of this document. Please e-mail: teradata-books@lists.ncr.com Any comments or materials (collectively referred to as “Feedback”) sent to NCR will be deemed non-confidential. NCR will have no obligation of any kind with respect to Feedback and will be free to use, reproduce, disclose, exhibit, display, transform, create derivative works of and distribute the Feedback and derivative works thereof without limitation on a royalty-free basis. Further, NCR will be free to use any ideas, concepts, know-how or techniques contained in such Feedback for any purpose whatsoever, including developing, manufacturing, or marketing products or services incorporating Feedback. Copyright © 2000 - 2006 by NCR Corporation. All Rights Reserved.SQL Reference: Fundamentals iii Preface Purpose SQL Reference: Fundamentals describes basic SQL data handling, SQL data definition, control, and manipulation, and the SQL lexicon. Use this book with the other books in the SQL Reference book set. Audience System administrators, database administrators, security administrators, application programmers, NCR field engineers, end users, and other technical personnel responsible for designing, maintaining, and using the Teradata Database will find this book useful. Experienced SQL users can also see simplified statement, data type, function, and expression descriptions in SQL/Data Dictionary Quick Reference. Supported Software Release This book supports Teradata ® Database V2R6.2. Prerequisites If you are not familiar with Teradata Database, you will find it useful to read Introduction to Teradata Warehouse before reading this book. You should be familiar with basic relational database management technology. This book is not an SQL primer.Preface Changes to This Book iv SQL Reference: Fundamentals Changes to This Book This book includes the following changes to support the current release. Date Description September 2006 • Added material to support BIGINT data type • Removed the restriction that the PARTITION BY option is not allowed in the CREATE JOIN INDEX statement for non-compressed join indexes • Removed the restriction that triggers cannot be defined on a table on which a join index is already defined • Updated the section on altering table structure and definition to indicate that ALTER TABLE can now be used to define, modify, or delete a COMPRESS attribute on an existing column • Updated Appendix E with new syntax for ALTER TABLE and CREATE TABLE • Moved the topics that identified valid and non-valid character ranges for KanjiEBCDIC, KanjiEUC, and KanjiShift-JIS object names from Chapter 2 to the International Character Set Support book May 2006 Removed RESTRICT from list of Teradata Database reserved words November 2005 • Added material to support new UDT and UDM feature • Added Appendix E, which details the differences in SQL between this release and previous releases • Removed the restriction that the PARTITION BY option is not allowed in the CREATE TABLE statement for global temporary tables and volatile tables November 2004 • Removed colons from stored procedure examples because colons are no longer required when local stored procedure variables or parameters are referenced in SQL statements • Added material to support new table function feature and new external stored procedure feature • Added overview of event processing using queue tables and the SELECT AND CONSUME statement • Removed the restriction that triggers cannot call stored procedures • Added material on new recursive query feature • Added material on new iterated requests feature • Added the restricted word list back into Appendix BPreface Additional Information SQL Reference: Fundamentals v Additional Information Additional information that supports this product and the Teradata Database is available at the following Web sites. Type of Information Description Source Overview of the release Information too late for the manuals The Release Definition provides the following information: • Overview of all the products in the release • Information received too late to be included in the manuals • Operating systems and Teradata Database versions that are certified to work with each product • Version numbers of each product and the documentation for each product • Information about available training and support center http://www.info.ncr.com/ Click General Search. In the Publication Product ID field, enter 1725 and click Search to bring up the following Release Definition: • Base System Release Definition B035-1725-096K Additional information related to this product Use the NCR Information Products Publishing Library site to view or download the most recent versions of all manuals. Specific manuals that supply related or additional information to this manual are listed. http://www.info.ncr.com/ Click General Search, and do one of the following: • In the Product Line field, select Software - Teradata Database for a list of all of the publications for this release, • In the Publication Product ID field, enter a book number. CD-ROM images This site contains a link to a downloadable CD-ROM image of all customer documentation for this release. Customers are authorized to create CDROMs for their use from this image. http://www.info.ncr.com/ Click General Search. In the Title or Keyword field, enter CD-ROM, and Click Search. Ordering information for manuals Use the NCR Information Products Publishing Library site to order printed versions of manuals. http://www.info.ncr.com/ Click How to Order under Print & CD Publications.Preface References to Microsoft Windows vi SQL Reference: Fundamentals References to Microsoft Windows This book refers to “Microsoft Windows.” For Teradata Database V2R6.2, such references mean Microsoft Windows Server 2003 32-bit and Microsoft Windows Server 2003 64-bit. General information about Teradata The Teradata home page provides links to numerous sources of information about Teradata. Links include: • Executive reports, case studies of customer experiences with Teradata, and thought leadership • Technical information, solutions, and expert advice • Press releases, mentions and media resources Teradata.com Type of Information Description SourceSQL Reference: Fundamentals vii Table of Contents Preface. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Supported Software Release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Changes to This Book. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv Additional Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v References to Microsoft Windows. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Chapter 1: Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Databases and Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Global Temporary Tables. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 Volatile Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 Columns. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Keys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Primary Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Secondary Indexes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Join Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Hash Indexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Referential Integrity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Views . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Triggers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Macros . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Stored Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 External Stored Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 User-Defined Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55Table of Contents viii SQL Reference: Fundamentals Roles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .57 User-Defined Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 Chapter 2: Basic SQL Syntax and Lexicon . . . . . . . . . . . . . . . . . . . . . . . .63 Structure of an SQL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63 SQL Lexicon Characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65 Keywords . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66 Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67 Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67 Standard Form for Data in Teradata Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71 Unqualified Object Names . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .73 Default Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75 Name Validation on Systems Enabled with Japanese Language Support . . . . . . . . . . . . . . . . .77 Object Name Translation and Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .81 Object Name Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .82 Finding the Internal Hexadecimal Representation for Object Names. . . . . . . . . . . . . . . . . . . .84 Specifying Names in a Logon String . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .86 Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87 NULL Keyword as a Literal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .90 Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .91 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .92 Delimiters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93 Separators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .94 Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95 Terminators. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 Null Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .98 Chapter 3: SQL Data Definition, Control, and Manipulation . .99 SQL Functional Families and Binding Styles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .99 Embedded SQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100 Data Definition Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .101 Altering Table Structure and Definition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103 Dropping and Renaming Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .104 Data Control Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .105Table of Contents SQL Reference: Fundamentals ix Data Manipulation Language . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Subqueries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Recursive Queries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 Query and Workload Analysis Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Help and Database Object Definition Tools. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Chapter 4: SQL Data Handling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Invoking SQL Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Transaction Processing in ANSI Session Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Transaction Processing in Teradata Session Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123 Multistatement Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Iterated Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Dynamic and Static SQL. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Dynamic SQL in Stored Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Using SELECT With Dynamic SQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 Event Processing Using Queue Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 Manipulating Nulls. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134 Session Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138 Session Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Return Codes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Statement Responses. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Success Response. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148 Warning Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Error Response (ANSI Session Mode Only). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Failure Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150 Chapter 5: Query Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Query Processing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Table Access. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Full-Table Scans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163 Collecting Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164Table of Contents x SQL Reference: Fundamentals Appendix A: Notation Conventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .167 Syntax Diagram Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .167 Character Shorthand Notation Used In This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .171 Predicate Calculus Notation Used in This Book . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .172 Appendix B: Restricted Words for V2R6.2. . . . . . . . . . . . . . . . . . . . . . .173 Reserved Words and Keywords for V2R6.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .173 Appendix C: Teradata Database Limits. . . . . . . . . . . . . . . . . . . . . . . . . . .203 System Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204 Database Limits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .206 Session Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .211 Appendix D: ANSI SQL Compliance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213 ANSI SQL Standard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .213 Terminology Differences Between ANSI SQL and Teradata . . . . . . . . . . . . . . . . . . . . . . . . . .216 SQL Flagger . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .217 Differences Between Teradata and ANSI SQL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .218 Appendix E: SQL Feature Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219 Notation Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219 Statements and Modifiers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .219 Data Types and Literals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .277 Functions, Operators, and Expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .280Table of Contents SQL Reference: Fundamentals xi Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289 Index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291Table of Contents xii SQL Reference: FundamentalsSQL Reference: Fundamentals 1 CHAPTER 1 Objects This chapter describes the objects you use to store, manage, and access data in the Teradata Database. Topics include: • Databases and Users • Tables • Columns • Data Types • Keys • Indexes • Views • Triggers • Macros • Stored Procedures and External Stored Procedures • User-Defined Functions • User-Defined Types (UDTs) and User-Defined Methods (UDMs) • Profiles • Roles Databases and Users Definitions A database is a collection of related tables, views, triggers, indexes, stored procedures, user-defined functions, and macros. A database also contains an allotment of space from which users can create and maintain their own objects, or other users or databases. A user is almost the same as a database, except that a user has a password and can log on to the system, whereas the database cannot. Defining Databases and Users Before you can create a database or user, you must have sufficient privileges granted to you. To create a database, use the CREATE DATABASE statement. You can specify the name of the database, the amount of storage to allocate, and other attributes.Chapter 1: Objects Tables 2 SQL Reference: Fundamentals To create a user, use the CREATE USER statement. The statement authorizes a new user identification (user name) for the database and specifies a password for user authentication. Because the system creates a database for each user, the CREATE USER statement is very similar to the CREATE DATABASE statement. Difference Between Users and Databases The difference between users and databases in the Teradata Database has important implications for matters related to access privileges, but neither the differences nor their implications are easy to understand. This is particularly true with respect to understanding fully the consequences of implicitly granted access privileges. Formally speaking, the difference between a user and a database is that a user has a password and a database does not. Users can also have default attributes such as time zone, date form, character set, role, and profile, while databases cannot. You might infer from this that databases are passive objects, while users are active objects. That is only true in the sense that databases cannot execute SQL statements. However, a query, macro, or stored procedure can execute using the privileges of the database. Tables Definitions A table is what is referred to in set theory terminology as a relation, from which the expression relational database is derived. Every relational table consists of one row of column headings (more commonly referred to as column names) and zero or more unique rows of data values. Formally speaking, each row represents what set theory calls a tuple. Each column represents what set theory calls an attribute. The number of rows (or tuples) in a table is referred to as its cardinality and the number of columns (or attributes) is referred to as its degree or arity. Defining Tables Use the CREATE TABLE statement to define base tables. The CREATE TABLE statement specifies a table name, one or more column names, and the attributes of each column. CREATE TABLE can also specify datablock size, percent freespace, and other physical attributes of the table. The CREATE/MODIFY USER and CREATE/MODIFY DATABASE statements provide options for creating permanent journal tables. Defining Indexes For a Table An index is a physical mechanism used to store and access the rows of a table. When you define a table, you can define a primary index and one or more secondary indexes. Chapter 1: Objects Tables SQL Reference: Fundamentals 3 All tables require a primary index. If you do not specify a column or set of columns as the primary index when you create a table, then CREATE TABLE specifies a primary index by default. For more information on indexes, see “Indexes” on page 17. Duplicate Rows in Tables Though both set theory and common sense prohibit duplicate rows in relational tables, the ANSI standard defines SQL based not on sets, but on bags, or multisets. A table defined not to permit duplicate rows is called a SET table because its properties are based on set theory, where set is defined as an unordered group of unique elements with no duplicates. A table defined to permit duplicate rows is called a MULTISET table because its properties are based on a multiset, or bag, model, where bag and multiset are defined as an unordered group of elements that may be duplicates. Temporary Tables Temporary tables are useful for temporary storage of data. Teradata Database supports three types of temporary tables. FOR more information on … SEE … rules for duplicate rows in a table CREATE TABLE in SQL Reference: Data Definition Statements. the result of an INSERT operation that would create a duplicate row INSERT in SQL Reference: Data Manipulation Statements. the result of an INSERT using a SELECT subquery that would create a duplicate row Type Usage Global temporary A global temporary table has a persistent table definition that is stored in the data dictionary. Any number of sessions can materialize and populate their own local copies that are retained until session logoff. Global temporary tables are useful for stroring temporary, intermediate results from multiple queries into working tables that are frequently used by applications. Global temporary tables are identical to ANSI global temporary tables. Volatile Like global temporary tables, the contents of volatile tables are only retained for the duration of a session. However, volatile tables do not have persistent definitions. To populate a volatile table, a session must first create the definition.Chapter 1: Objects Tables 4 SQL Reference: Fundamentals Materialized instances of a global temporary table share the following characteristics with volatile tables: • Private to the session that created them. • Contents cannot be shared by other sessions. • Optionally emptied at the end of each transaction using the ON COMMIT PRESERVE/DELETE rows option in the CREATE TABLE statement. • Activity optionally logged in the transient journal using the LOG/NO LOG option in the CREATE TABLE statement. • Dropped automatically when a session ends. For details about the individual characteristics of global temporary and volatile tables, see “Global Temporary Tables” on page 5 and “Volatile Tables” on page 9. Queue Tables Teradata Database supports queue tables, which are similar to ordinary base tables, with the additional unique property of behaving like an asynchronous first-in-first-out (FIFO) queue. Queue tables are useful for applications that want to submit queries that wait for data to be inserted into queue tables without polling. When you create a queue table, you must define a TIMESTAMP column with a default value of CURRENT_TIMESTAMP. The values in the column indicate the time the rows were inserted into the queue table, unless different, user-supplied values are inserted. You can then use a SELECT AND CONSUME statement, which operates like a FIFO pop: • Data is returned from the row with the oldest timestamp value in the specified queue table. • The row is deleted from the queue table, guaranteeing that the row is processed only once. If no rows are available, the transaction enters a delay state until one of the following occurs: • A row is inserted into the queue table. • The transaction aborts, either as a result of direct user intervention, such as the ABORT statement, or indirect user intervention, such as a DROP TABLE statement on the queue table. To perform a FIFO peek on a queue table, use a SELECT statement. Global temporary trace Global temporary trace tables are useful for debugging external routines (UDFs, UDMs, and external stored procedures). During execution, external routines can write trace output to columns in a global temporary trace table. Like global temporary tables, global temporary trace tables have persistent definitions, but do not retain rows across sessions. Type UsageChapter 1: Objects Global Temporary Tables SQL Reference: Fundamentals 5 Global Temporary Tables Introduction Global temporary tables allow you to define a table template in the database schema, providing large savings for applications that require well known temporary table definitions. The definition for a global temporary table is persistent and stored in the data dictionary. Space usage is charged to login user temporary space. Each user session can materialize as many as 2000 global temporary tables at a time. How Global Temporary Tables Work To create the base definition for a global temporary table, use the CREATE TABLE statement and specify the keywords GLOBAL TEMPORARY to describe the table type. Once created, the table exists only as a definition. It has no rows and no physical instantiation. When any application in a session accesses a table with the same name as the defined base table, and the table has not already been materialized in that session, then that table is materialized as a real relation using the stored definition. Because that initial invocation is generally due to an INSERT statement, a temporary table—in the strictest sense—is usually populated immediately upon its materialization. There are only two occasions when an empty global temporary table is materialized: • A CREATE INDEX statement is issued on the table. • A COLLECT STATISTICS statement is issued on the table. The following table summarizes this information. Note: Issuing a SELECT, UPDATE, or DELETE on a global temporary table that is not materialized produces the same result as issuing a SELECT, UPDATE, or DELETE on an empty global temporary table that is materialized. WHEN this statement is issued on a global temporary table that has not yet been materialized … THEN a local instance of the global temporary table is materialized and it is … INSERT populated with data upon its materialization. CREATE INDEX … ON TEMPORARY … COLLECT STATISTICS … ON TEMPORARY … not populated with data upon its materialization.Chapter 1: Objects Global Temporary Tables 6 SQL Reference: Fundamentals Example For example, suppose there are four sessions, Session 1, Session 2, Session 3, and Session 4 and two users, User_1 and User_2. Consider the scenario in the following two tables. Step Session … Does this … The result is this … 1 1 The DBA creates a global temporary table definition in the database scheme named globdb.gt1 according to the following CREATE TABLE statement: CREATE GLOBAL TEMPORARY TABLE globdb.gt1, LOG (f1 INT NOT NULL PRIMARY KEY, f2 DATE, f3 FLOAT) ON COMMIT PRESERVE ROWS; The global temporary table definition is created and stored in the database schema. 2 1 User_1 logs on an SQL session and references globdb.gt1 using the following INSERT statement: INSERT globdb.gt1 (1, 980101, 11.1); Session 1 creates a local instance of the global temporary table definition globdb.gt1. This is also referred to as a materialized temporary table. Immediately upon materialization, the table is populated with a single row having the following values. f1=1 f2=980101 f3=11.1 This means that the contents of this local instance of the global temporary table definition is not empty when it is created. From this point on, any INSERT/DELETE/UPDATE statement that references globdb.gt1 in Session 1 is mapped to this local instance of the table. 3 2 User_2 logs on an SQL session and issues the following SELECT statement. SELECT * FROM globdb.gt1; No rows are returned because Session 2 has not yet materialized a local instance of globdb.gt1.Chapter 1: Objects Global Temporary Tables SQL Reference: Fundamentals 7 User_1 and User_2 continue their work, logging onto two additional sessions as described in the following table. 4 2 User_2 issues the following INSERT statement: INSERT globdb.gt1 (2, 980202, 22.2); Session 2 creates a local instance of the global temporary table definition globdb.gt1. The table is populated, immediately upon materialization, with a single row having the following values. f1=2 f2=980202 f3=22.2 From this point on, any INSERT/DELETE/UPDATE statement that references globdb.gt1 in Session 2 is mapped to this local instance of the table. 5 2 User_2 logs again issues the following SELECT statement: SELECT * FROM globdb.gt1; A single row containing the data (2, 980202, 22.2) is returned to the application. 6 1 User_1 logs off from Session 1. The local instance of globdb.gt1 for Session 1 is dropped. 7 2 User_2 logs off from Session 2. The local instance of globdb.gt1 for Session 2 is dropped. Step Session … Does this … The result is this … 1 3 User_1 logs on another SQL session 3 and issues the following SELECT statement: SELECT * FROM globdb.gt1; No rows are returned because Session 3 has not yet materialized a local instance of globdb.gt1. 2 3 User_1 issues the following INSERT statement: INSERT globdb.gt1 (3, 980303, 33.3); Session 3 created a local instance of the global temporary table definition globdb.gt1. The table is populated, immediately upon materialization, with a single row having the following values. f1=3 f2=980303 f3=33.3 From this point on, any INSERT/DELETE/UPDATE statement that references globdb.gt1 in Session 3 maps to this local instance of the table. Step Session … Does this … The result is this …Chapter 1: Objects Global Temporary Tables 8 SQL Reference: Fundamentals With the exception of a few options (see “CREATE TABLE” in SQL Reference: Data Definition Statements for an explanation of the features not available for global temporary base tables), materialized temporary tables have the same properties as permanent tables. After a global temporary table definition is materialized in a session, all further references to the table are made to the materialized table. No additional copies of the base definition are materialized for the session. This global temporary table is defined for exclusive use by the session whose application materialized it. 3 3 User_1 again issues the following SELECT statement: SELECT * FROM globdb.gt1; A single row containing the data (3, 980303, 33.3) is returned to the application. 4 4 User_2 logs on Session 4 and issues the following CREATE INDEX statement: CREATE INDEX (f2) ON TEMPORARY globdb.gt1; An empty local global temporary table named globdb.gt1 is created for Session 4. This is one of only two cases in which a local instance of a global temporary table is materialized without data. The other would be a COLLECT STATISTICS statement—in this case, the following statement: COLLECT STATISTICS ON TEMPORARY globdb.gt1; 5 4 User_2 issues the following SELECT statement: SELECT * FROM globdb.gt1; No rows are returned because the local instance of globdb.gt1 for Session 4 is empty. 6 4 User_2 issues the following SHOW TABLE statement: SHOW TABLE globdb.gt1; CREATE SET GLOBAL TEMPORARY TABLE globdb.gt1, FALLBACK, LOG ( f1 INTEGER NOT NULL, f2 DATE FORMAT 'YYYY-MM-DD', f3 FLOAT) UNIQUE PRIMARY INDEX (f1) ON COMMIT PRESERVE ROWS; 7 4 User_2 issues the following SHOW TEMPORARY TABLE statement: SHOW TEMPORARY TABLE globdb.gt1; CREATE SET GLOBAL TEMPORARY TABLE globdb.gt1, FALLBACK, LOG ( f1 INTEGER NOT NULL, f2 DATE FORMAT 'YYYY-MM-DD', f3 FLOAT) UNIQUE PRIMARY INDEX (f1) INDEX (f2) ON COMMIT PRESERVE ROWS; Note that this report indicates the new index f2 that has been created for the local instance of the temporary table. Step Session … Does this … The result is this …Chapter 1: Objects Volatile Tables SQL Reference: Fundamentals 9 Materialized global temporary tables differ from permanent tables in the following ways: • They are always empty when first materialized. • Their contents cannot be shared by another session. • The contents can optionally be emptied at the end of each transaction. • The materialized table is dropped automatically at the end of each session. Limitations You cannot use the following CREATE TABLE options for global temporary tables: • WITH DATA • Permanent journaling • Referential integrity constraints This means that a temporary table cannot be the referencing or referenced table in a referential integrity constraint. References to global temporary tables are not permitted in FastLoad, MultiLoad, or FastExport. Archive, Restore, and TableRebuild operate on base global temporary tables only. Non-ANSI Extensions Transient journaling options on the global temporary table definition are permitted using the CREATE TABLE statement. You can modify the transient journaling and ON COMMIT options for base global temporary tables using the ALTER TABLE statement. Privileges Required To materialize a global temporary table, you must have the appropriate privilege on the base global temporary table or on the containing database or user as required by the statement that materializes the table. No access logging is performed on materialized global temporary tables, so no access log entries are generated. Volatile Tables Creating Volatile Tables Neither the definition nor the contents of a volatile table persist across a system restart. You must use CREATE TABLE with the VOLATILE keyword to create a new volatile table each time you start a session in which it is needed. Chapter 1: Objects Volatile Tables 10 SQL Reference: Fundamentals What this means is that you can create volatile tables as you need them. Being able to create a table quickly provides you with the ability to build scratch tables whenever you need them. Any volatile tables you create are dropped automatically as soon as your session logs off. Volatile tables are always created in the login user space, regardless of the current default database setting. That is, the database name for the table is the login user name. Space usage is charged to login user spool space. Each user session can materialize as many as 1000 volatile tables at a time. Limitations The following CREATE TABLE options are not permitted for volatile tables: • Permanent journaling • Referential integrity constraints This means that a volatile table cannot be the referencing or referenced table in a referential integrity constraint. • Check constraints • Compressed columns • DEFAULT clause • TITLE clause • Named indexes References to volatile tables are not permitted in FastLoad or MultiLoad. For more information, see “CREATE TABLE” in SQL Reference: Data Definition Statements. Non-ANSI Extensions Volatile tables are not defined in ANSI. Privileges Required To create a volatile table, you do not need any privileges. No access logging is performed on volatile tables, so no access log entries are generated. Volatile Table Maintenence Among Multiple Sessions Volatile tables are private to a session. This means that you can log on multiple sessions and create volatile tables with the same name in each session. However, at the time you create a volatile table, the name must be unique among all global and permanent temporary table names in the database that has the name of the login user.Chapter 1: Objects Volatile Tables SQL Reference: Fundamentals 11 For example, suppose you log on two sessions, Session 1 and Session 2. Assume the default database name is your login user name. Consider the following scenario. Stage In Session 1, you … In Session 2, you … The result is this … 1 Create a volatile table named VT1. Create a volatile table named VT1. Each session creates its own copy of volatile table VT1 using your login user name as the database. 2 Create a permanent table with an unqualified table name of VT2. Session 1 creates a permanent table named VT2 using your login user name as the database. 3 Create a volatile table named VT2. Session 2 receives a CREATE TABLE error, because there is already a permanent table with that name. 4 Create a volatile table named VT3. Session 1 creates a volatile table named VT3 using your login user name as the database. 5 Create a permanent table with an unqualified table name of VT3. Session 2 creates a permanent table named VT3 using your login user name as the database. Because a volatile table is known only to the session that creates it, a permanent table with the same name as the volatile table VT3 in Session 1 can be created as a permanent table in Session 2. 6 Insert into VT3. Session 1 references volatile table VT3. Note: Volatile tables take precedence over permanent tables in the same database in a session. Because Session 1 has a volatile table VT3, any reference to VT3 in Session 1 is mapped to the volatile table VT3 until it is dropped (see Step 10). On the other hand, in Session 2, references to VT3 remain mapped to the permanent table named VT3. 7 Create volatile table VT3. Session 2 receives a CREATE TABLE error for attempting to create the volatile table VT3 because of the existence of that permanent table. 8 Insert into VT3. Session 2 references permanent table VT3. 9 Drop VT3. Session 2 drops volatile table VT3. 10 Select from VT3. Session 1 references the permanent table VT3.Chapter 1: Objects Columns 12 SQL Reference: Fundamentals Columns Definition A column is a structural component of a table and has a name and a declared type. Each row in a table has exactly one value for each column. Each value in a row is a value in the declared type of the column. The declared type includes nulls and values of the declared type. A column value is the smallest unit of data that can be selected from or updated for a table. Defining Columns The column definition clause of the CREATE TABLE statement defines the table column elements. A name and a data type must be specified for each column defined for a table. Each column can be further defined with one or more attribute definitions. Here is an example that creates a table called employee with three columns: CREATE TABLE employee (deptno INTEGER ,name CHARACTER(23) ,hiredate DATE); The following optional subclauses are also elements of the SQL column definition clause: • Data type attribute declaration, such as NOT NULL, FORMAT, and TITLE • COMPRESS column storage attributes clause • Column constraint attributes clause, such as PRIMARY KEY • UNIQUE table-level definition clause • REFERENCES table-level definition clause • CHECK constraint table-level definition clause Related Topics FOR more information on … SEE … data types “Data Types” on page 13. CREATE TABLE and the column definition clause SQL Reference: Data Definition Statements.Chapter 1: Objects Data Types SQL Reference: Fundamentals 13 Data Types Introduction Every data value belongs to an SQL data type. For example, when you define a column in a CREATE TABLE statement, you must specify the data type of the column. The set of data values that a column defines can belong to one of the following data types: Numeric Data Types A numeric value is either an exact numeric number (integer or decimal) or an approximate numeric number (floating point). Use the following SQL data types to specify numeric values. Character Data Types Character data types represent characters that belong to a given character set. Use the following SQL data types to specify character data. • Numeric • Character • Datetime • Interval • Byte • UDT Type Description BIGINT Represents a signed, binary integer value from -9,223,372,036,854,775,808 to 9,223,372,036,854,775,807. INTEGER Represents a signed, binary integer value from -2,147,483,648 to 2,147,483,647. SMALLINT Represents a signed binary integer value in the range -32768 to 32767. BYTEINT Represents a signed binary integer value in the range -128 to 127. REAL Represent a value in sign/magnitude form. DOUBLE PRECISION FLOAT DECIMAL [(n[,m])] Represent a decimal number of n digits, with m of those n digits to the right of the decimal point. NUMERIC [(n[,m])] Type Description CHAR Represents a fixed length character string for Teradata Database internal character storage. VARCHAR(n) Represents a variable length character string of length n for Teradata Database internal character storage.Chapter 1: Objects Data Types 14 SQL Reference: Fundamentals DateTime Data Types DateTime values represent dates, times, and timestamps. Use the following SQL data types to specify DateTime values. Interval Data Types An interval value is a span of time. There are two mutually exclusive interval type categories. LONG VARCHAR LONG VARCHAR specifies the longest permissible variable length character string for Teradata Database internal character storage. CLOB Represents a large character string. A character large object (CLOB) column can store character data, such as simple text, HTML, or XML documents. Type Description Type Description DATE Represents a date value that includes year, month, and day components. TIME Represents a time value that includes hour, minute, second, and fractional second components. TIMESTAMP Represents a timestamp value that includes year, month, day, hour, minute, second, and fractional second components. TIME WITH TIME ZONE Represents a time value that includes hour, minute, second, fractional second, and time zone components. TIMESTAMP WITH TIIME ZONE Represents a timestamp value that includes year, month, day, hour, minute, second, fractional second, and time zone components. Category Type Description Year-Month • INTERVAL YEAR • INTERVAL YEAR TO MONTH • INTERVAL MONTH Represent a time span that can include a number of years and months. Day-Time • INTERVAL DAY • INTERVAL DAY TO HOUR • INTERVAL DAY TO MINUTE • INTERVAL DAY TO SECOND • INTERVAL HOUR • INTERVAL HOUR TO MINUTE • INTERVAL HOUR TO SECOND • INTERVAL MINUTE • INTERVAL MINUTE TO SECOND • INTERVAL SECOND Represent a time span that can include a number of days, hours, minutes, or seconds.Chapter 1: Objects Data Types SQL Reference: Fundamentals 15 Byte Data Types Byte data types store raw data as logical bit streams. For any machine, BYTE, VARBYTE, and BLOB data is transmitted directly from the memory of the client system. BLOB is ANSI SQL-2003-compliant. BYTE and VARBYTE are Teradata extensions to the ANSI SQL-2003 standard. UDT Data Types UDT data types are custom data types that you define with the CREATE TYPE statement. Teradata Database supports distinct and structured UDTs. For more details on UDTs, including a synopsis of the steps you take to develop and use UDTs, see “User-Defined Types” on page 58. Related Topics For detailed information on data types, see SQL Reference: Data Types and Literals. Type Description BYTE Represents a fixed-length binary string. VARBYTE Represents a variable-length binary string. BLOB Represents a large binary string of raw bytes. A binary large object (BLOB) column can store binary objects, such as graphics, video clips, files, and documents. Type Description Distinct A UDT that is based on a single predefined data type, such as INTEGER or VARCHAR. Structured A UDT that is a collection of one or more fields called attributes, each of which is defined as a predefined data type or other UDT (which allows nesting).Chapter 1: Objects Keys 16 SQL Reference: Fundamentals Keys Definitions Keys and Referential Integrity Teradata Database uses primary and foreign keys to maintain referential integrity. For additional information, see “Referential Integrity” on page 36. Effect on Row Distribution Because Teradata Database uses a unique primary or secondary index to enforce a primary key, the primary key can affect how Teradata Database distributes and retrieves rows. For more information, see “Primary Indexes” on page 22 and “Secondary Indexes” on page 25. Differences Between Primary Keys and Primary Indexes The following table summarizes the differences between keys and indexes using the primary key and primary index for purposes of comparison. Term Definition Primary Key A primary key is a column, or combination of columns, in a table that uniquely identifies each row in the table. The values defining a primary key for a table: • Must be unique • Cannot change • Cannot be null Foreign Key A foreign key is a column, or combination of columns, in a table that is also the primary key in one or more additional tables in the same database. Foreign keys provide a mechanism to link related tables based on key values. Primary Key Primary Index Important element of logical data model. Not used in logical data model. Used to maintain referential integrity. Used to distribute and retrieve data. Must be unique to identify each row. Can be unique or nonunique. Values cannot change. Values can change. Cannot be null. Can be null. Does not imply access path. Defines the most common access path. Not required for physical table definition. Required for physical table definition.Chapter 1: Objects Indexes SQL Reference: Fundamentals 17 Indexes Definition An index is a mechanism that the SQL query optimizer can use to make table access more performant. Indexes enhance data access by providing a more-or-less direct path to stored data to avoid performing full table scans to locate the small number of rows you typically want to retrieve or update. The Teradata Database parallel architecture makes indexing an aid to better performance, not a crutch necessary to ensure adequate performance. Full table scans are not something to be feared in the Teradata Database environment. This means that the sorts of unplanned, ad hoc queries that characterize the data warehouse process, and that often are not supported by indexes, perform very effectively for Teradata Database using full table scans. The classic index for a relational database is itself a file made up of rows having two parts: • A (possibly unique) data field in the referenced table. • A pointer to the location of that row in the base table (if the index is unique) or a pointer to all possible locations of rows with that data field value (if the index is nonunique). Because the Teradata Database is a massively parallel architecture, it requires a more efficient means of distributing and retrieving its data. One such method is hashing. All Teradata Database indexes are based on row hash values rather than raw table column values, even though secondary, hash, and join indexes can be stored in order of their values to make them more useful for satisfying range conditions. Selectivity of Indexes An index that retrieves many rows is said to have weak selectivity. An index that retrieves few rows is said to be strongly selective. The more strongly selective an index is, the more useful it is. In some cases, it is possible to link together several weakly selective nonunique secondary indexes by bit mapping them. The result is effectively a strongly selective index and a dramatic reduction in the number of table rows that must be accessed. For more information on linking weakly selective secondary indexes into a strongly selective unit using bit mapping, see “NUSI Bit Mapping” on page 28. Row Hash and RowID Teradata Database table rows are self-indexing with respect to their primary index and so require no additional storage space. When a row is inserted into a table, the relational database manager stores the 32-bit row hash value of the primary index with it. Because row hash values are not necessarily unique, the relational database manager also generates a unique 32-bit numeric value (called the Uniqueness Value) that it appends to the row hash value, forming a unique RowID. This RowID makes each row in a table uniquely identifiable and ensures that hash collisions do not occur.Chapter 1: Objects Indexes 18 SQL Reference: Fundamentals If a table is defined with a partitioned primary index (PPI), the RowID also includes the partition number to which the row was assigned. For more information on PPIs, see “Partitioned and Non-Partitioned Primary Indexes” on page 20. The first row having a specific row hash value is always assigned a uniqueness value of 1, which becomes the highest current uniqueness value. Thereafter, each time another row having the same row hash value is inserted, the row is assigned the current high value incremented by 1, and that value becomes the current high value. Table rows having the same row hash value are stored on disk sorted in the ascending order of RowID. Uniqueness values are not reused except for the special case in which the highest valued row within a row hash is deleted from a table. A RowID for a row might change, for instance, when a primary index or partitioning column is changed, or when there is complex update of the table. Index Hash Mapping Rows are distributed across the AMPS using a hashing algorithm that computes a row hash value based on the primary index. The row hash is a 32-bit value. The higher-order 16 bits of a hash value determine an associated hash bucket. Teradata Database databases have 65536 hash buckets. The hash buckets are distributed as evenly as possible among the AMPs on a system. Teradata Database maintains a hash map—an index of which hash buckets live on which AMPs—that it uses to determine whether rows belong to an AMP based on their row hash values. Row assignment is performed in a manner that ensures as equal a distribution as possible among all the AMPs on a system. Advantages of Indexes The intent of indexes is to lessen the time it takes to retrieve rows from a database. The faster the retrieval, the better. Disadvantages of Indexes Perhaps not so obvious is the disadvantage of using indexes. • They must be updated every time a row is updated, deleted, or added to a table. This is only a consideration for indexes other than the primary index in the Teradata Database environment. The more indexes you have defined for a table, the bigger the potential update downside becomes. Because of this, secondary, join, and hash indexes are rarely appropriate for OLTP situations. • All Teradata Database secondary indexes are stored in subtables, and join and hash indexes are stored in separate tables, exerting a burden on system storage space.Chapter 1: Objects Indexes SQL Reference: Fundamentals 19 • When FALLBACK is defined for a table, a further storage space burden is created because secondary index subtables are always duplicated whenever FALLBACK is defined for a table. An additional burden on system storage space is exerted when FALLBACK is defined for join indexes or hash indexes or both. For this reason, it is extremely important to use the EXPLAIN modifier to determine optimum data manipulation statement syntax and index usage before putting statements and indexes to work in a production environment. For more information on EXPLAIN, see SQL Reference: Data Manipulation Statements. Teradata Database Index Types Teradata Database provides four different index types: • Primary index All Teradata Database tables require a primary index because the system distributes tables on their primary indexes. Primary indexes can be: • Unique or nonunique • Partitioned or non-partitioned • Secondary index Secondary indexes can be unique or nonunique. • Join index (JI) • Hash index Unique Indexes A unique index, like a primary key, has a unique value for each row in a table. Teradata Database defines two different types of unique index. • Unique primary index (UPI) UPIs provide optimal data distribution and are typically assigned to the primary key for a table. When a NUPI makes better sense for a table, then the primary key is frequently assigned to be a USI. • Unique secondary index (USI) USIs guarantee that each complete index value is unique, while ensuring that data access based on it is always a two-AMP operation. Nonunique Indexes A nonunique index does not require its values to be unique. There are occasions when a nonunique index is the best choice as the primary index for a table. NUSIs are also very useful for many decision support situations.Chapter 1: Objects Indexes 20 SQL Reference: Fundamentals Partitioned and Non-Partitioned Primary Indexes Primary indexes can be partitioned or non-partitioned. A non-partitioned primary index (NPPI) is the traditional primary index by which rows are assigned to AMPs. A partitioned primary index (PPI) allows rows to be partitioned, based on some set of columns, on the AMP to which they are distributed, and ordered by the hash of the primary index columns within the partition. A PPI can be used to improve query performance through partition elimination. A PPI provides a useful alternative to an NPPI for executing range queries against a table, while still providing efficient join and aggregation strategies on the primary index. Join Indexes A join index is an indexing structure containing columns from one or more base tables and is generally used to resolve queries and eliminate the need to access and join the base tables it represents. Teradata Database join indexes can be defined in the following general ways. • Simple or aggregate • Single- or multitable • Hash-ordered or value-ordered • Complete or sparse For details, see “Join Indexes” on page 30. Hash Indexes Hash indexes are used for the same purposes as are single-table join indexes, and are less complicated to define. However, a join index offers more choices. For additional information, see “Hash Indexes” on page 34. Creating Indexes For a Table Use the CREATE TABLE statement to define a primary index and one or more secondary indexes. You can define the primary index (and any secondary index) as unique, depending on whether duplicate values are to be allowed in the indexed column set. A partitioned primary index cannot be defined as unique if one or more partitioning columns are not included in the primary index. To create hash or join indexes, use the CREATE HASH INDEX and CREATE JOIN INDEX statements, respectively.Chapter 1: Objects Indexes SQL Reference: Fundamentals 21 Using EXPLAIN and Teradata Index Wizard to Determine the Usefulness of Indexes One important thing to remember is that the use of indexes by the optimizer is not under user control in a relational database management system. That is, the only references made to indexes in the SQL language concern their definition and not their use. The SQL data manipulation language statements do not provide for any specification of indexes. There are several implications of this behavior. • First, it is very important to collect statistics regularly to ensure that the optimizer has access to current information about how to best optimize any query or update made to the database. For additional information concerning collecting and maintaining accurate database statistics, see “COLLECT STATISTICS” in SQL Reference: Data Definition Statements. • Second, it is even more important to build your queries and updates in such a way that you know their performance will be most optimal. Apart from good logical database design, one way to ensure that you are accessing your data in the most efficient manner possible is to use the EXPLAIN modifier to try out various candidate queries or updates and to note which indexes are used by the optimizer in their execution (if any) as well as examining the relative length of time required to complete the operation. There are several methods you can use to determine optimal sets of secondary indexes tailored to particular application workloads: • Teradata Index Wizard • EXPLAIN reports The Teradata Index Wizard client utility provides a method of determining optimum secondary indexes for a given SQL statement workload automatically and then verifying that the proposed indexes actually produce the expected performance enhancements. See the following references for more information about the Teradata Index Wizard: • Teradata Index Wizard User Guide • SQL Reference: Statement and Transaction Processing You can produce and analyze EXPLAIN reports using either the Teradata Visual Explain client utility or the SQL EXPLAIN request modifier. For each statement in the request, EXPLAIN output provides you with the following basic information: • The step-by-step access method the optimizer would use to execute the specified data manipulation statement given the current set of table statistics it has to work with. • The relative time it would take to perform the data manipulation statement. While you cannot rely on the reported statement execution time as an absolute, you can rely on it as a relative means for comparison with other candidate data manipulation statements against the same tables with the same statistics defined.Chapter 1: Objects Primary Indexes 22 SQL Reference: Fundamentals Primary Indexes Introduction The primary index for a table controls the distribution and retrieval of the data for that table across the AMPs. Both distribution and retrieval of the data is controlled using the Teradata Database hashing algorithm (see “Row Hash and RowID” on page 17 and “Index Hash Mapping” on page 18). If the primary index is defined as a partitioned primary index (PPI), the data is partitioned, based on some set of columns, on each AMP, and ordered by the hash of the primary index columns within the partition. Data accessed based on a primary index is always a one-AMP operation because a row and its index are stored on the same AMP. This is true whether the primary index is unique or nonunique, and whether it is partitioned or non-partitioned. Tables Require a Primary Index All Teradata Database tables require a primary index. To create a primary index, use the CREATE TABLE statement. If you do not assign a primary index explicitly when you create a table, Teradata Database assigns a primary index, based on the following rules. FOR more information on … SEE … using the EXPLAIN request modifier SQL Reference: Data Manipulation Statements using the Teradata Visual Explain client utility Teradata Visual Explain User Guide additional performance-related information about how to use the access and join plan reports produced by EXPLAIN to optimize the performance of your databases • Database Design • Performance Management WHEN a CREATE TABLE statement defines a … THEN Teradata Database selects the … Primary Index Primary Key Unique Column Constraint No Yes No primary key column set to be a UPI. No No Yes first column or columns having a UNIQUE constraint to be a UPI. No Yes Yes primary key column set to be a UPI.Chapter 1: Objects Primary Indexes SQL Reference: Fundamentals 23 In general, the best practice is to specify a primary index instead of having Teradata Database select a default primary index. Uniform Distribution of Data and Optimal Access Considerations When choosing the primary index for a table, there are two essential factors to keep in mind: uniform distribution of the data and optimal access. With respect to uniform data distribution, consider the following factors: • The more distinct the primary index values, the better. • Rows having the same primary index value are distributed to the same AMP. • Parallel processing is more efficient when table rows are distributed evenly across the AMPs. With respect to optimal data access, consider the following factors: • Choose the primary index on the most frequently used access path. For example, if rows are generally accessed by a range query, consider defining a partitioned primary index on the table that creates a useful set of partitions. If the table is frequently joined with a specific set of tables, consider defining the primary index on the column set that is typically used as the join condition. • Primary index operations must provide the full primary index value. • Primary index retrievals on a single value are always one-AMP operations. While it is true that the columns you choose to be the primary index for a table are often the same columns that define the primary key, it is also true that primary indexes often comprise fields that are neither unique nor components of the primary key for the table. Unique and Nonunique Primary Index Considerations In addition to uniform distribution of data and optimal access considerations, other guidelines and performance considerations apply to selecting a unique or a nonunique column set as the primary index for a table. No No No first column defined for the table to be a NUPI. If the data type of the first column in the table is UDT or LOB, then the CREATE TABLE operation aborts and the system returns an error message. WHEN a CREATE TABLE statement defines a … THEN Teradata Database selects the … Primary Index Primary Key Unique Column ConstraintChapter 1: Objects Primary Indexes 24 SQL Reference: Fundamentals Generally, other considerations can include the following: • Primary and other alternate key column sets • The value range seen when using predicates in a WHERE clause • Whether access can involve multiple rows or a spool file or both For more information on criteria for selecting a primary index, see Database Design. Partitioning Considerations The decision to define a Partitioned Primary Index (PPI) for a table depends on how its rows are most frequently accessed. PPIs are designed to optimize range queries while also providing efficient primary index join strategies. For range queries, only rows of the qualified partitions need to be accessed. PPI increases query efficiency by avoiding full table scans without the overhead and maintenance costs of secondary indexes. Various partitioning strategies are possible: • For some applications, defining the partitions such that each has approximately the same number of rows might be an effective strategy. • For other applications, it might be desirable to have a varying number of rows per partition. For example, more frequently accessed data (such as for the current year) might be divided into finer partitions (such as weeks) but other data (such as previous years) may have coarser partitions (such as months or multiples of months). • Alternatively, it might be important to define each range with equal width, even if the number of rows per range varies. The most important factors for PPIs are accessibility and maximization of partition elimination. In all cases, it is critical for parallel efficiency to define a primary index that distributes the rows of the table fairly evenly across the AMPs. For more information on partitioning considerations, see Database Design. Primary Index Summary Teradata Database primary indexes have the following properties. • Defined with the CREATE TABLE data definition statement. CREATE INDEX is used only to create secondary indexes. • Modified with the ALTER TABLE data definition statement. Some modifications, such as partitioning and primary index columns, require an empty table. • Automatically assigned by CREATE TABLE if you do not explicitly define a primary index. However, the best practice is to always specify the primary index, because the default may not be appropriate for the table. • Can be composed of as many as 64 columns. • A maximum of one can be defined per table.Chapter 1: Objects Secondary Indexes SQL Reference: Fundamentals 25 • Can be partitioned or non-partitioned. Partitioned primary indexes are not automatically assigned. You must explicitly define a partitioned primary index. • Can be unique or non-unique. Note that a partitioned primary index can only be unique if all the partitioning columns are also included as primary index columns. If the primary index does not include all the partitioning columns, uniqueness on the primary index columns may be enforced with a unique secondary index on the same columns as the primary index. • Defined as non-unique if the primary index is not defined explicitly as unique or if the primary index is specified for a single column SET table. • Controls data distribution and retrieval using the Teradata hashing algorithm. • Improves performance when used correctly in the WHERE clause of an SQL data manipulation statement to perform the following actions. • Single-AMP retrievals • Joins between tables with identical primary indexes, the optimal scenario • Partition elimination when the primary index is partitioned Related Topics Consult the following books for more detailed information on using primary indexes to enhance the performance of your databases: • Database Design • Performance Management Secondary Indexes Introduction Secondary indexes are never required for Teradata Database tables, but they can often improve system performance. You create secondary indexes explicitly using the CREATE TABLE and CREATE INDEX statements. Teradata Database can implicitly create unique secondary indexes; for example, when you use a CREATE TABLE statement that specifies a primary index, Teradata Database implicitly creates unique secondary indexes on column sets that you specify using PRIMARY KEY or UNIQUE constraints. Creating a secondary index causes the Teradata Database to build a separate internal subtable to contain the index rows, thus adding another set of rows that requires updating each time a table row is inserted, deleted, or updated. Nonunique secondary indexes (NUSIs) can be specified as either hash-ordered or value-ordered. Value-ordered NUSIs are limited to a single numeric-valued (including DATE) sort key whose size is four or fewer bytes.Chapter 1: Objects Secondary Indexes 26 SQL Reference: Fundamentals Secondary index subtables are also duplicated whenever a table is defined with FALLBACK. After the table is created and usage patterns have developed, additional secondary indexes can be defined with the CREATE INDEX statement. Differences Between Unique and Nonunique Secondary Indexes Teradata Database processes USIs and NUSIs very differently. Consider the following statements that define a USI and a NUSI. The following table highlights differences in the build process for the preceding statements. Secondary Index Statement USI CREATE UNIQUE INDEX (customer_number) ON customer_table; NUSI CREATE INDEX (customer_name) ON customer_table; USI Build Process NUSI Build Process Each AMP accesses its subset of the base table rows. Each AMP accesses its subset of the base table rows. Each AMP copies the secondary index value and appends the RowID for the base table row. Each AMP builds a spool file containing each secondary index value found followed by the RowID for the row it came from. Each AMP creates a Row Hash on the secondary index value and puts all three values onto the BYNET. For hash-ordered NUSIs, each AMP sorts the RowIDs for each secondary index value into ascending order. For value-ordered NUSIs, the rows are sorted by NUSI value order. The appropriate AMP receives the data and creates a row in the index subtable. If the AMP receives a row with a duplicate index value, an error is reported. For hash-ordered NUSIs, each AMP creates a row hash value for each secondary index value on a local basis and creates a row in its portion of the index subtable. For value-ordered NUSIs, storage is based on NUSI value rather than the row hash value for the secondary index. Each row contains one or more RowIDs for the index value.Chapter 1: Objects Secondary Indexes SQL Reference: Fundamentals 27 Consider the following statements that access a USI and a NUSI. The following table identifies differences for the access process of the preceding statements. Note: The NUSI is not used if the estimated number of rows to be read in the base table is equal to or greater than the estimated number of data blocks in the base table; in this case, a full table scan is done, or, if appropriate, partition scans are done. NUSIs and Covering The Optimizer aggressively pursues NUSIs when they cover a query. Covered columns can be specified anywhere in the query, including the select list, the WHERE clause, aggregate functions, GROUP BY clauses, expressions, and so on. Presence of a WHERE condition on each indexed column is not a prerequisite for using a NUSI to cover a query. Value-Ordered NUSIs Value-ordered NUSIs are very efficient for range conditions, and more so when strongly selective or when combined with covering. Because the NUSI rows are sorted by data value, it is possible to search only a portion of the index subtable for a given range of key values. Secondary Index Statement USI SELECT * FROM customer_table WHERE customer_number=12; NUSI SELECT * FROM customer_table WHERE customer_name = 'SMITH'; USI Access Process NUSI Access Process The supplied index value hashes to the corresponding secondary index row. A message containing the secondary index value is broadcast to every AMP. The retrieved base table RowID is used to access the specific data row. For a hash-ordered NUSI, each AMP creates a local row hash and uses it to access its portion of the index subtable to see if a corresponding row exists. Value-ordered NUSI index subtable values are scanned only for the range of values specified by the query. The process is complete. This is typically a two-AMP operation. If an index row is found, the AMP uses the RowID or value order list to access the corresponding base table rows. The process is complete. This is always an all-AMP operation, with the exception of a NUSI that is defined on the same columns as the primary index.Chapter 1: Objects Secondary Indexes 28 SQL Reference: Fundamentals Value-ordered NUSIs have the following limitations. • The sort key is limited to a single numeric or DATE column. • The sort key column must be four or fewer bytes. The following query is an example of the sort of SELECT statement for which value-ordered NUSIs were designed. SELECT * FROM Orders WHERE o_date BETWEEN DATE '1998-10-01' AND DATE '1998-10-07'; Multiple Secondary Indexes and Composites Database designers frequently define multiple secondary indexes on a table. For example, the following statements define two secondary indexes on the EMPLOYEE table: CREATE INDEX (department_number) ON EMPLOYEE; CREATE INDEX (job_code) ON EMPLOYEE; The WHERE clause in the following query specifies the columns that have the secondary indexes defined on them: SELECT last_name, first_name, salary_amount FROM employee WHERE department_number = 500 AND job_code = 2147; Whether the Optimizer chooses to include one, all, or none of the secondary indexes in its query plan depends entirely on their individual and composite selectivity. NUSI Bit Mapping Bit mapping is a technique used by the Optimizer to effectively link several weakly selective indexes in a way that creates a result that drastically reduces the number of base rows that must be accessed to retrieve the desired data. The process determines common rowIDs among multiple NUSI values by means of the logical intersection operation. Bit mapping is significantly faster than the three-part process of copying, sorting, and comparing rowID lists. Additionally, the technique dramatically reduces the number of base table I/Os required to retrieve the requested rows. FOR more information on … SEE … multiple secondary index access Database Design composite secondary index access other aspects of index selectionChapter 1: Objects Secondary Indexes SQL Reference: Fundamentals 29 Secondary Index Summary Teradata SQL secondary indexes have the following properties. • Can enhance the speed of data retrieval. Because of this, secondary indexes are most useful in decision support applications. • Do not affect data distribution. • Can be a maximum of 32 defined per table. • Can be composed of as many as 64 columns. • For a value-ordered NUSI, only a single numeric or DATE column of four or fewer bytes may be specified for the sort key. • For a hash-ordered covering index, only a single column may be specified for the hash ordering. • Can be created or dropped dynamically as data usage changes or if they are found not to be useful for optimizing data retrieval performance. • Require additional disk space to store subtables. • Require additional I/Os on inserts and deletes. Because of this, secondary indexes might not be as useful in OLTP applications. • Should not be defined on columns whose values change frequently. • Should not include columns that do not enhance selectivity. • Should not use composite secondary indexes when multiple single column indexes and bit mapping might be used instead. • Composite secondary indexes is useful if it reduces the number of rows that must be accessed. • The Optimizer does not use composite secondary indexes unless there are explicit values for each column in the index. • Most efficient for selecting a small number of rows. • Can be unique or non-unique. • NUSIs can be hash-ordered r value-ordered, and can optionally include covering columns. • Cannot be partitioned, but can be defined on a table with a partitioned primary index. FOR more information on … SEE … when Teradata Database performs NUSI bit mapping Database Design how NUSI bit maps are computed using the EXPLAIN modifier to determine if bit mapping is being used for your indexes • Database Design • SQL Reference: Data Manipulation StatementsChapter 1: Objects Join Indexes 30 SQL Reference: Fundamentals Summary of USI and NUSI Properties Unique and nonunique secondary indexes have the following properties. For More Information About Secondary Indexes See “SQL Data Definition Language Statement Syntax” of SQL Reference: Data Definition Statements under “CREATE TABLE” and “CREATE INDEX” for more information. Also consult the following manuals for more detailed information on using secondary indexes to enhance the performance of your databases: • Database Design • Performance Management Join Indexes Introduction Join indexes are not indexes in the usual sense of the word. They are file structures designed to permit queries (join queries in the case of multitable join indexes) to be resolved by accessing the index instead of having to access and join their underlying base tables. You can use join indexes to: • Define a prejoin table on frequently joined columns (with optional aggregation) without denormalizing the database. • Create a full or partial replication of a base table with a primary index on a foreign key column table to facilitate joins of very large tables by hashing their rows to the same AMP as the large table. • Define a summary table without denormalizing the database. You can define a join index on one or several tables. Depending on how the index is defined, join indexes can also be useful for queries where the index structure contains only some of the columns referenced in the statement. This situation is referred to as a partial cover of the query. Unlike traditional indexes, join indexes do not implicitly store pointers to their associated base table rows. Instead, they are generally used as a fast path final access point that eliminates the USI NUSI • Guarantee that each complete index value is unique. • Any access using the index is a two-AMP operation. • Useful for locating rows having a specific value in the index. • Can be hash-ordered or value-ordered. Value-ordered NUSIs are particularly useful for enhancing the performance of range queries. • Can include covering columns. • Any access using the index is an all-AMP operation.Chapter 1: Objects Join Indexes SQL Reference: Fundamentals 31 need to access and join the base tables they represent. They substitute for rather than point to base table rows. The only exception to this is the case where an index partially covers a query. If the index is defined using either the ROWID keyword or the UPI or USI of its base table as one of its columns, then it can be used to join with the base table to cover the query. Defining Join Indexes To create a join index, use the CREATE JOIN INDEX statement. For example, suppose that a common task is to look up customer orders by customer number and date. You might create a join index like the following, linking the customer table, the order table, and the order detail table: CREATE JOIN INDEX cust_ord2 AS SELECT cust.customerid,cust.loc,ord.ordid,item,qty,odate FROM cust, ord, orditm WHERE cust.customerid = ord.customerid AND ord.ordid = orditm.ordid; Multitable Join Indexes A multitable join index stores and maintains the joined rows of two or more tables and, optionally, aggregates selected columns. Multitable join indexes are for join queries that are performed frequently enough to justify defining a prejoin on the joined columns. A multitable join index is useful for queries where the index structure contains all the columns referenced by one or more joins, thereby allowing the index to cover that part of the query, making it possible to retrieve the requested data from the index rather than accessing its underlying base tables. For obvious reasons, an index with this property is often referred to as a covering index. Single-Table Join Indexes Single-table join indexes are very useful for resolving joins on large tables without having to redistribute the joined rows across the AMPs. Single-table join indexes facilitate joins by hashing a frequently joined subset of base table columns to the same AMP as the table rows to which they are frequently joined. This enhanced geography eliminates BYNET traffic as well as often providing a smaller sized row to be read and joined. Aggregate Join Indexes When query performance is of utmost importance, aggregate join indexes offer an extremely efficient, cost-effective method of resolving queries that frequently specify the same aggregate operations on the same column or columns. When aggregate join indexes are available, the system does not have to repeat aggregate calculations for every query.Chapter 1: Objects Join Indexes 32 SQL Reference: Fundamentals You can define an aggregate join index on two or more tables, or on a single table. A single-table aggregate join index includes a summary table with: • A subset of columns from a base table • Additional columns for the aggregate summaries of the base table columns Sparse Join Indexes You can create join indexes that limit the number of rows in the index to only those that are accessed when, for example, a frequently run query references only a small, well known subset of the rows of a large base table. By using a constant expression to filter the rows included in the join index, you can create what is known as a sparse index. Any join index, whether simple or aggregate, multitable or single-table, can be sparse. To create a sparse index, use the WHERE clause in the CREATE JOIN INDEX statement. Effects of Join Indexes Join index limits affect the following Teradata Database functions and features. • Load Utilities MultiLoad and FastLoad utilities cannot be used to load or unload data into base tables that have a join index defined on them because join indexes are not maintained during the execution of these utilities. If an error occurs because of the join index, drop the join index and recreate it after loading data into that table. The TPump utility, which performs standard SQL row inserts and updates, can be used to load or unload data into base tables with join indexes because it properly maintains join indexes during execution. However, in some cases, performance may improve by dropping join indexes on the table prior to the load and recreating them after the load. • ARC (Archive and Recovery) Archive and Recovery cannot be used on a join index itself. Archiving is permitted on a base table or database that has an associated join index defined. Before a restore of such a base table or database, you must drop the existing join index definition. Before using any such index again in the execution of queries, you must recreate the join index definition. • Permanent Journal Recovery Using a permanent journal to recover a base table (that is, ROLLBACK or ROLLFORWARD) with an associated join index defined is permitted. The join index is not automatically rebuilt during the recovery process. Instead, it is marked as non-valid and it must be dropped and recreated before it can be used again in the execution of queries.Chapter 1: Objects Join Indexes SQL Reference: Fundamentals 33 Comparison of Join Indexes and Base Tables In most respects, a join index is similar to a base table. For example, you can do the following things to a join index: • Create nonunique secondary indexes on its columns. • Execute COLLECT STATISTICS, DROP STATISTICS, HELP, and SHOW statements. • Partition its primary index, if it is a non-compressed join index. Note: Unlike a base table that has a PPI, however, you cannot use COLLECT STATISTICS to collect PARTITION statistics on a non-compressed join index that has a PPI. Unlike base tables, you cannot do the following things with join indexes: • Query or update join index rows explicitly. • Store and maintain arbitrary query results such as expressions. Note: You can maintain aggregates or sparse indexes if you define the join index to do so. • Create explicit unique indexes on its columns. Related Topics FOR more information on … SEE … creating join indexes “CREATE JOIN INDEX” in SQL Reference: Data Definition Statements dropping join indexes “DROP JOIN INDEX” in SQL Reference: Data Definition Statements displaying the attributes of the columns defined by a join index “HELP JOIN INDEX” in SQL Reference: Data Definition Statements using join indexes to enhance the performance of your databases • Database Design • Performance Management • SQL Reference: Data Definition Statements • database design considerations for join indexes • improving join index performance Database DesignChapter 1: Objects Hash Indexes 34 SQL Reference: Fundamentals Hash Indexes Introduction Hash indexes are used for the same purposes as single-table join indexes. The following table lists the principal differences between hash indexes and single-table join indexes. Hash indexes are useful for creating a full or partial replication of a base table with a primary index on a foreign key column to facilitate joins of very large tables by hashing them to the same AMP. You can define a hash index on one table only. The functionality of hash indexes is a subset to that of single-table join indexes. Comparison of Hash and Single-Table Join Indexes The reasons for using hash indexes are similar to those for using single-table join indexes. Not only can hash indexes optionally be specified to be distributed in such a way that their rows are AMP-local with their associated base table rows, they also implicitly provide an alternate direct access path to those base table rows. This facility makes hash indexes somewhat similar to secondary indexes in function. Hash indexes are also useful for covering queries so that the base table need not be accessed at all. Hash Index Single-Table Join Index Column list cannot contain aggregate or ordered analytical functions. Column list can contain aggregate functions. Cannot have a secondary index. Can have a secondary index. Supports transparently added, system-defined columns that point to the underlying base table rows. Does not implicitly add underlying base table row pointers. Pointers to underlying base table rows can be created explicitly by defining one element of the column list using the ROWID keyword or the UPI or USI of the base table. FOR information on … SEE … using CREATE HASH INDEX to create a hash index SQL Reference: Data Definition Statements using DROP HASH INDEX to drop a hash index using HELP HASH INDEX to display the data types of the columns defined by a hash index database design considerations for hash indexes Database DesignChapter 1: Objects Hash Indexes SQL Reference: Fundamentals 35 The following list summarizes the similarities of hash and single-table join indexes: • Primary function of both is to improve query performance. • Both are maintained automatically by the system when the relevant columns of their base table are updated by a DELETE, INSERT, UPDATE, or MERGE statement. • Both can be the object of any of the following SQL statements: • COLLECT STATISTICS • DROP STATISTICS • HELP INDEX • SHOW • Both receive their space allocation from permanent space and are stored in distinct tables. • The storage organization for both supports a compressed format to reduce storage space, but for a hash index, Teradata Database makes this decision. • Both can be FALLBACK protected. • Neither can be queried or directly updated. • Neither can store an arbitrary query result. • Both share the same restrictions for use with the MultiLoad, FastLoad, and Archive/Recovery utilities. • A hash index implicitly defines a direct access path to base table rows. A join index may be explicitly specified to define a direct access path to base table rows. Effects of Hash Indexes Join index limits affect the following Teradata Database functions and features. • ARC (Archive and Recovery) Archive and Recovery cannot be used on a hash index itself. Archiving is permitted on a base table or database that has an associated hash index defined. During a restore of such a base table or database, the system does not rebuild the hash index. You must drop the existing hash index definition and create a new one before any such index can be used again in the execution of queries. • Load Utilities MultiLoad and FastLoad utilities cannot be used to load or unload data into base tables that have an associated hash index defined on them because hash indexes are not maintained during the execution of these utilities. The hash index must be dropped and recreated after that table has been loaded. The TPump utility, which performs standard SQL row inserts and updates, can be used because hash indexes are properly maintained during its execution. However, in some cases, performance may improve by dropping hash indexes on the table prior to the load and recreating them after the load. • Permanent Journal Recovery Using a permanent journal to recover a base table using ROLLBACK or ROLLFORWARD with an associated hash index defined is permitted. The hash index is not automatically Chapter 1: Objects Referential Integrity 36 SQL Reference: Fundamentals rebuilt during the recovery process. Instead, the hash index is marked as non-valid and it must be dropped and recreated before it can be used again in the execution of queries. Queries Using a Hash Index In most respects, a hash index is similar to a base table. For example, you can perform COLLECT STATISTICS, DROP STATISTICS, HELP, and SHOW statements on a hash index. Unlike base tables, you cannot do the following things with hash indexes: • Query or update hash index rows explicitly. • Store and maintain arbitrary query results such as expressions. • Create explicit unique indexes on its columns. • Partition the primary index of the hash index. For More Information About Hash Indexes Consult the following manuals for more detailed information on using hash indexes to enhance the performance of your databases: • Database Design • Performance Management • SQL Reference: Data Definition Statements Referential Integrity Introduction Referential integrity (RI) is defined as all the following notions. • The concept of relationships between tables, based on the definition of a primary key (or UNIQUE alternate key) and a foreign key. • A mechanism that provides for specification of columns within a referencing table that are foreign keys for columns in some other referenced table. Referenced columns must be defined as one of the following. • Primary key columns • Unique columns • A reliable mechanism for preventing accidental database corruption when performing inserts, updates, and deletes. Referential integrity requires that a row having a non-null value for a referencing column cannot exist in a table if an equal value does not exist in a referenced column.Chapter 1: Objects Referential Integrity SQL Reference: Fundamentals 37 Varieties of Referential Integrity Enforcement Supported by Teradata Database Teradata Database supports two forms of declarative SQL for enforcing referential integrity: • A standard method that enforces RI on a row-by-row basis • A batch method that enforces RI on a statement basis Both methods offer the same measure of integrity enforcement, but perform it in different ways. A third form is related to these because it provides a declarative definition for a referential relationship, but it does not enforce that relationship. Enforcement of the declared referential relationship is left to the user by any appropriate method. Referencing (Child) Table The referencing table is referred to as the child table, and the specified child table columns are the referencing columns. Note: Referencing columns must have the same numbers and types of columns, data types, and sensitivity as the referenced table keys. COMPRESS is not allowed on either referenced or referencing columns and column-level constraints are not compared. Referenced (Parent) Table A child table must have a parent, and the referenced table is referred to as the parent table. The parent key columns in the parent table are the referenced columns. Because the referenced columns are defined as unique constraints, they must be one of the following unique indexes. • A unique primary index (UPI), defined as NOT NULL • A unique secondary index (USI), defined as NOT NULL Terms Related to Referential Integrity The following terms are used to explain the concept of referential integrity. Term Definition Child Table A table where the referential constraints are defined. Child table and referencing table are synonyms. Parent Table The table referenced by a child table. Parent table and referenced table are synonyms. Primary Key A unique identifier for a row of a table. UNIQUE Alternate KeyChapter 1: Objects Referential Integrity 38 SQL Reference: Fundamentals Why Referential Integrity Is Important Consider the employee and payroll tables for any business. With referential integrity constraints, the two tables work together as one. When one table gets updated, the other table also gets updated. The following case depicts a useful referential integrity scenario. Looking for a better career, Mr. Clark Johnson leaves his company. Clark Johnson is deleted from the employee table. The payroll table, however, does not get updated because the payroll clerk simply forgets to do so. Consequently, Mr. Clark Johnson keeps getting paid. With good database design, referential integrity relationship would have been defined on these tables. They would have been linked and, depending on the defined constraints, the deletion of Clark Johnson from the employee table could not be performed unless it was accompanied by the deletion of Clark Johnson from the payroll table. Foreign Key A column set in the child table that is also the primary key (or a UNIQUE alternate key) in the parent table. Foreign keys can consist of as many as 64 different columns. Referential Constraint A constraint defined on a column set or a table to ensure referential integrity. For example, consider the following table definition: CREATE TABLE A (A1 CHAR(10) REFERENCES B (B1), A2 INTEGER FOREIGN KEY (A1,A2) REFERENCES C PRIMARY INDEX (A1)); This CREATE TABLE statement specifies the following referential integrity constraints. This constraint … Is defined at this level … 1 column. Implicit foreign key A1 references the parent key B1 in table B. 2 table. Explicit composite foreign key (A1, A2) implicitly references the UPI (or a USI) of parent table C, which must be two columns, the first typed CHAR(10) and the second typed INTEGER. Both parent table columns must also be defined as NOT NULL. Term DefinitionChapter 1: Objects Referential Integrity SQL Reference: Fundamentals 39 Besides data integrity and data consistency, referential integrity also has the benefits listed in the following table. Rules for Assigning Columns as FOREIGN KEYS The FOREIGN KEY columns in the referencing table must be identical in definition with the keys in the referenced table. Corresponding columns must have the same data type and case sensitivity. • The COMPRESS option is not permitted on either the referenced or referencing column(s). • Column level constraints are not compared. • A one-column FOREIGN KEY cannot reference a single column in a multi-column primary or unique key—the foreign and primary/unique key must contain the same number of columns. Circular References Are Allowed References can be defined as circular in that TableA can reference TableB, which can reference TableA. In this case, at least one set of FOREIGN KEYS must be defined on nullable columns. If the FOREIGN KEYS in TableA are on columns defined as nullable, then rows could be inserted into TableA with nulls for the FOREIGN KEY columns. Once the appropriate rows exist in TableB, the nulls of the FOREIGN KEY columns in TableA could then be updated to contain non-null values which match the TableB values. References Can Be to the Table Itself FOREIGN KEY references can also be to the same table that contains the FOREIGN KEY. The referenced columns must be different columns than the FOREIGN KEY, and both the referenced and referencing columns must subscribe to the referential integrity rules. Benefit Description Increases development productivity It is not necessary to code SQL statements to enforce referential constraints. The Teradata Database automatically enforces referential integrity. Requires fewer programs to be written All update activities are programmed to ensure that referential constraints are not violated. The Teradata Database enforces referential integrity in all environments. No additional programs are required. Improves performance The Teradata Database chooses the most efficient method to enforce the referential constraints. The Teradata Database can optimize queries based on the fact that there is referential integrity.Chapter 1: Objects Referential Integrity 40 SQL Reference: Fundamentals CREATE and ALTER TABLE Syntax Referential integrity affects the syntax and semantics of CREATE TABLE and ALTER TABLE. For more details, see “ALTER TABLE” and “CREATE TABLE” in SQL Reference: Data Definition Statements. Maintaining Foreign Keys Definition of a FOREIGN KEY requires that the Teradata Database maintain the integrity defined between the referenced and referencing table. The Teradata Database maintains the integrity of foreign keys as explained in the following table. FOR this data manipulation activity … The system verifies that … A row is inserted into a referencing table and foreign key columns are defined to be NOT NULL. a row exists in the referenced table with the same values as those in the foreign key columns. If such a row does not exist, then an error is returned. If the foreign key contains multiple columns, and if any one column value of the foreign key is null, then none of the foreign key values are validated. The values in foreign key columns are altered to be NOT NULL. a row exists in the referenced table that contains values equal to the altered values of all of the foreign key columns. If such a row does not exist, then an error is returned. A row is deleted from a referenced table. no rows exist in referencing tables with foreign key values equal to those of the row to be deleted. If such rows exist, then an error is returned. Before a referenced column in a referenced table is updated. no rows exist in a referencing table with foreign key values equal to those of the referenced columns. If such rows exist, then an error is returned. Before the structure of columns defined as foreign keys or referenced by foreign keys is altered. the change would not violate the rules for definition of a foreign key constraint. An ALTER TABLE or DROP INDEX statement attempting to change such a columns structure returns an error. A table referenced by another is dropped. the referencing table has dropped its foreign key reference to the referenced table.Chapter 1: Objects Referential Integrity SQL Reference: Fundamentals 41 Referential Integrity and the ARC Utility The Archive (ARC) utility archives and restores individual tables. It also copies tables from one database to another. When a table is restored or copied into a database, the dictionary definition of that table is also restored. The dictionary definitions of both the referenced (parent) and referencing (child) table contain the complete definition of a reference. By restoring a single table, it is possible to create an inconsistent reference definition in the Teradata Database. When either a parent or child table is restored, the reference is marked as inconsistent in the dictionary definitions. The ARC utility can validate these references once the restore is done. An ALTER TABLE statement adds a foreign key reference to a table. The same processes occur whether the reference is defined for standard or for soft referential integrity. all of the values in the foreign key columns are validated against columns in the referenced table. When the system parses ALTER TABLE, it defines an error table that: • Has the same columns and primary index as the target table of the ALTER TABLE statement. • Has a name that is the same as the target table name suffixed with the reference index number. A reference index number is assigned to each foreign key constraint for a table. To determine the number, use one of the following system views. • RI_Child_Tables • RI_Distinct_Children • RI_Distinct_Parents • RI_Parent_Tables • Is created under the same user or database as the table being altered. If a table already exists with the same name as that generated for the error table then an error is returned to the ALTER TABLE statement. Rows in the referencing table that contain values in the foreign key columns that cannot be found in any row of the referenced table are copied into the error table (the base data of the target table is not modified). It is your responsibility to: • Correct data values in the referenced or referencing tables so that full referential integrity exists between the two tables. Use the rows in the error table to define which corrections to make. • Maintain the error table. FOR this data manipulation activity … The system verifies that …Chapter 1: Objects Views 42 SQL Reference: Fundamentals While a table is marked as inconsistent, no updates, inserts, or deletes are permitted. The table is fully usable only when the inconsistencies are resolved (see below). This restriction is true for both hard and soft (Referential Constraint) referential integrity constraints. It is possible that the user either intends to or must revert to a definition of a table which results in an inconsistent reference on that table. The Archive and Restore operations are the most common cause of such inconsistencies. To remove inconsistent references from a child table that is archived and restored, follow these steps: 1 After archiving the child table, drop the parent table. 2 Restore the child table. When the child table is restored, the parent table no longer exists. The normal ALTER TABLE DROP FOREIGN KEY statement does not work, because the parent table references cannot be resolved. 3 Use the DROP INCONSISTENT REFERENCES option to remove these inconsistent references from a table. The syntax is: ALTER TABLE database_name.table_name DROP INCONSISTENT REFERENCES You must have DROP privileges on the target table of the statement to perform this option, which removes all inconsistent internal indexes used to establish references. For further information, see Teradata Archive/Recovery Utility Reference or Teradata ASF2 Tape Reader User Guide. Referential Integrity and the FastLoad and MultiLoad Utilities Foreign key references are not supported for any table that is the target table for a FastLoad or MultiLoad. For further details, see: • Database Design • Teradata FastLoad Reference • Teradata MultiLoad Reference Views Views and Tables A view can be compared to a window through which you can see selected portions of a database. Views are used to retrieve portions of one or more tables or other views. Views look like tables to a user, but they are virtual, not physical, tables. They display data in columns and rows and, in general, can be used as if they were physical tables. However, only the column definitions for a view are stored: views are not physical tables.Chapter 1: Objects Views SQL Reference: Fundamentals 43 A view does not contain data: it is a virtual table whose definition is stored in the data dictionary. The view is not materialized until it is referenced by a statement. Some operations that are permitted for the manipulation of tables are not valid for views, and other operations are restricted, depending on the view definition. Defining a View The CREATE VIEW statement defines a view. The statement names the view and its columns, defines a SELECT on one or more columns from one or more underlying tables and/or views, and can include conditional expressions and aggregate operators to limit the row retrieval. Why Use Views? The primary reason to use views is to simplify end user access to the Teradata database. Views provide a constant vantage point from which to examine and manipulate the database. Their perspective is altered neither by adding or nor by dropping columns from its component base tables unless those columns are part of the view definition. From an administrative perspective, views are useful for providing an easily maintained level of security and authorization. For example, users in a Human Resources department can access tables containing sensitive payroll information without being able to see salary and bonus columns. Views also provide administrators with an ability to control read and update privileges on the database with little effort. Restrictions on Views Some operations that are permitted on base tables are not permitted on views—sometimes for obvious reasons and sometimes not. The following set of rules outlines the restrictions on how views can be created and used. • You cannot create an index on a view. • A view definition cannot contain an ORDER BY clause. • Any derived columns in a view must explicitly specify view column names, for example by using an AS clause or by providing a column list immediately after the view name. • You cannot update tables from a view under the following circumstances: • The view is defined as a join view (defined on more than one table) • The view contains derived columns. • The view definition contains a DISTINCT clause. • The view definition contains a GROUP BY clause. • The view defines the same column more than once.Chapter 1: Objects Triggers 44 SQL Reference: Fundamentals Triggers Definition Triggers are active database objects associated with a subject table. A trigger essentially consists of a stored SQL statement or a block of SQL statements. Triggers execute when an INSERT, UPDATE, DELETE, or MERGE modifies a specified column or columns in the subject table. Typically, a stored trigger performs an UPDATE, INSERT, DELETE, MERGE, or other SQL operation on one or more tables, which may possibly include the subject table. Triggers in Teradata Database conform to the ANSI SQL-2003 standard, and also provide some additional features. Triggers have two types of granularity: • Row triggers fire once for each row of the subject table that is changed by the triggering event and that satisfies any qualifying condition included in the row trigger definition. • Statement triggers fire once upon the execution of the triggering statement. You can create, alter, and drop triggers. For details on creating, dropping, and altering triggers, see SQL Reference: Data Definition Statements. Process Flow for a Trigger The general process flow for a trigger is as follows. Note that this is a logical flow, not a physical re-enactment of how the Teradata Database processes a trigger. 1 The triggering event occurs on the subject table. 2 A determination is made as to whether triggers defined on the subject table are to become active upon a triggering event. 3 Qualified triggers are examined to determine the trigger action time, whether they are defined to fire before or after the triggering event. IF you want to … THEN use … define a trigger CREATE TRIGGER. • enable a trigger • disable a trigger • change the creation timestamp for a trigger ALTER TRIGGER. Disabling a trigger stops the trigger from functioning, but leaves the trigger definition in place as an object. This allows utility operations on a table that are not permitted on tables with enabled triggers. Enabling a trigger restores its active state. remove a trigger from the system permanently DROP TRIGGER.Chapter 1: Objects Triggers SQL Reference: Fundamentals 45 4 When multiple triggers qualify, then they fire normally in the ANSI-specified order of creation timestamp. To override the creation timestamp and specify a different execution order of triggers, you can use the ORDER clause, a Teradata extension. Even if triggers are created without the ORDER clause, you can redefine the order of execution by changing the trigger creation timestamp using the ALTER TRIGGER statement. 5 The triggered SQL statements (triggered action) execute. If the trigger definition uses a REFERENCING clause to specify that old, new, or both old and new data for the triggered action is to be collected under a correlation name (an alias), then that information is stored in transition tables or transition rows as follows: • OLD [ROW] or NEW [ROW] values, or both, under old (or new) values correlation name. • Entire set of rows as OLD TABLE or NEW TABLE under old (or new) values table alias. 6 The trigger passes control to the next trigger, if defined, in a cascaded sequence. The sequence can include recursive triggers. Otherwise, control passes to the next statement in the application. 7 If any of the actions involved in the triggering event or the triggered actions abort, then all of the actions are aborted. Restrictions on Using Triggers Most Teradata load utilities cannot access a table that has an active trigger. An application that uses triggers can use ALTER TRIGGER to disable the trigger and enable the load. The application must be sure that loading a table with disabled triggers does not result in a mismatch in a user defined relationship with a table referenced in the triggered action. The other restrictions on triggers include: • BEFORE statement triggers are not allowed. • BEFORE triggers cannot have data-changing statements as triggered action (triggered SQL statements). • BEFORE triggers cannot access OLD TABLE and NEW TABLE. • Triggers and hash indexes are mutually exclusive. You cannot define triggers on a table on which a hash index is already defined. • A positioned (updatable cursor) UPDATE or DELETE is not allowed to fire a trigger. An attempt to do so generates an error.Chapter 1: Objects Macros 46 SQL Reference: Fundamentals Related Topics Macros Introduction A frequently used SQL statement or series of statements can be incorporated into a macro and defined using the SQL CREATE MACRO statement. See “CREATE MACRO” in SQL Reference: Data Definition Statements. The statements in the macro are performed using the EXECUTE statement. See “EXECUTE (Macro Form)” in SQL Reference: Data Manipulation Statements. A macro can include an EXECUTE statement that executes another macro. Definition A macro consists of one or more statements that can be executed by performing a single statement. Each time the macro is performed, one or more rows of data can be returned. Performing a macro is similar to performing a multistatement request (see “Multistatement Requests” on page 121). Single-User and Multiuser Macros You can create a macro for your own use, or grant execution authorization to others. For example, your macro might enable a user in another department to perform operations on the data in the Teradata Database. When executing the macro, a user need not be aware of the database being accessed, the tables affected, or even the results. FOR detailed information on … SEE … • guidelines for creating triggers • conditions that cause triggers to fire • trigger action that occurs when a trigger fires • the trigger action time • when to use row triggers and when to use statement triggers CREATE TRIGGER in SQL Reference: Data Definition Statements. • temporarily disabling triggers • enabling triggers • changing the creation timestamp of a trigger ALTER TRIGGER in SQL Reference: Data Definition Statements. permanently removing triggers from the system DROP TRIGGER in SQL Reference: Data Definition Statements.Chapter 1: Objects Macros SQL Reference: Fundamentals 47 Multistatement Transactions Versus Macros Although you can enter a multistatement operation interactively using an explicit transaction (either BT/ET or COMMIT), a better practice is to define such an operation as a macro because an explicit transaction holds locks placed on objects by statements in the transaction until the statement sequence is completed with an END TRANSACTION or COMMIT statement. If you were to enter such a sequence interactively from BTEQ, items in the database would be locked to others while you typed and entered each statement. Contents of a Macro With the exception of CREATE AUTHORIZATION and REPLACE AUTHORIZATION, a data definition statement is allowed in macro if it is the only SQL statement in that macro. A data definition statement is not resolved until the macro is executed, at which time unqualified database object references are fully resolved using the default database of the user submitting the EXECUTE statement. If this is not the desired result, you must fully qualify all object references in a data definition statement in the macro body. A macro can contain parameters that are substituted with data values each time the macro is executed. It also can include a USING modifier, which allows the parameters to be filled with data from an external source such as a disk file. A COLON character prefixes references to a parameter name in the macro. Parameters cannot be used for data object names. Executing a Macro Regardless of the number of statements in a macro, the Teradata Database treats it as a single request. When you execute a macro, either all its statements are processed successfully or none are processed. If a macro fails, it is aborted, any updates are backed out, and the database is returned to its original state. Ways to Perform SQL Macros in Embedded SQL Macros in an embedded SQL program are performed in one of the following ways. IF the macro … THEN use … is a single statement, and that statement returns no data • the EXEC statement to specify static execution of the macro -or- • the PREPARE and EXECUTE statements to specify dynamic execution. Use DESCRIBE to verify that the single statement of the macro is not a data returning statement. • consists of multiple statements • returns data a cursor, either static or dynamic. The type of cursor used depends on the specific macro and on the needs of the application.Chapter 1: Objects Stored Procedures 48 SQL Reference: Fundamentals Static SQL Macro Execution in Embedded SQL Static SQL macro execution is associated with a macro cursor using the macro form of the DECLARE CURSOR statement. When you perform a static macro, you must use the EXEC form to distinguish it from the dynamic SQL statement EXECUTE. Dynamic SQL Macro Execution in Embedded SQL Define dynamic macro execution using the PREPARE statement with the statement string containing an EXEC macro_name statement rather than a single-statement request. The dynamic request is then associated with a dynamic cursor. See “DECLARE CURSOR (Macro Form)” in SQL Reference: Data Manipulation Statements for further information on the use of macros. Dropping, Replacing, Renaming, and Retrieving Information About a Macro For more information, see SQL Reference: Data Definition Statements. Stored Procedures Introduction Stored procedures are called Persistent Stored Modules in the ANSI SQL-2003 standard. They are written in SQL and consist of a set of control and condition handling statements that make SQL a computationally complete programming language. These features provide a server-based procedural interface to the Teradata Database for application programmers. Teradata stored procedure facilities are a subset of and conform to the ANSI SQL-2003 standards for semantics. IF you want to … THEN use the following statement … drop a macro DROP MACRO redefine an existing macro REPLACE MACRO rename a macro RENAME MACRO get the attributes for a macro HELP MACRO get the data definition statement most recently used to create, replace, or modify a macro SHOW MACROChapter 1: Objects Stored Procedures SQL Reference: Fundamentals 49 Elements of Stored Procedures The set of statements constituting the main tasks of the stored procedure is called the stored procedure body, which can consist of a single statement or a compound statement, or block. A single statement stored procedure body can contain one control statement, such as LOOP or WHILE, or one SQL DDL, DML, or DCL statement, including dynamic SQL. Some statements are not allowed, including: • Any declaration (local variable, cursor, or condition handler) statement • A cursor statement (OPEN, FETCH, or CLOSE) A compound statement stored procedure body consists of a BEGIN-END statement enclosing a set of declarations and statements, including: • Local variable declarations • Cursor declarations • Condition handler declaration statements • Control statements • SQL DML, DDL, and DCL statements supported by stored procedures, including dynamic SQL Compound statements can also be nested. For information about control statements, parameters, local variables, and labels, see SQL Reference: Stored Procedures and Embedded SQL. Privileges for Stored Procedures The security for stored procedures is similar to that for other Teradata database objects like tables, macros, views, and triggers. The rights to ALTER PROCEDURE, CREATE PROCEDURE, DROP PROCEDURE, and EXECUTE PROCEDURE can be granted using the GRANT statement and revoked using the REVOKE statement. Of these: • CREATE PROCEDURE is only a database-level privilege. • ALTER PROCEDURE, DROP PROCEDURE and EXECUTE PROCEDURE privileges can be granted at the object level and database or user level. • Only DROP PROCEDURE is an automatic privilege for all users. This is granted when a new user or database is created. • EXECUTE PROCEDURE is an automatic privilege only for the creator of a stored procedure, granted at the time of creation. Chapter 1: Objects Stored Procedures 50 SQL Reference: Fundamentals Creating Stored Procedures A stored procedure can be created from: • BTEQ utility using the COMPILE command • CLIv2 applications, ODBC, JDBC, and Teradata SQL Assistant (formerly called Queryman) using the SQL CREATE PROCEDURE or REPLACE PROCEDURE statement. The procedures are stored in the user database space as objects and are executed on the server. For the syntax of data definition statements related to stored procedures, including CREATE PROCEDURE and REPLACE PROCEDURE, see SQL Reference: Data Definition Statements. Note: The stored procedure definitions in the next examples are designed only to demonstrate the usage of the feature. They are not recommended for use. Example Assume you want to define a stored procedure NewProc to add new employees to the Employee table and retrieve the name of the department to which the employee belongs. You can also report an error, in case the row that you are trying to insert already exists, and handle that error condition. The CREATE PROCEDURE statement looks like this: CREATE PROCEDURE NewProc (IN name CHAR(12), IN number INTEGER, IN dept INTEGER, OUT dname CHAR(10), INOUT errstr VARCHAR(30)) BEGIN DECLARE CONTINUE HANDLER FOR SQLSTATE VALUE '23505' SET errstr = 'Duplicate Row.'; INSERT INTO Employee (EmpName, EmpNo, DeptNo ) VALUES (name, number, dept); SELECT DeptName INTO dname FROM Department WHERE DeptNo = dept; END; This stored procedure defines parameters that must be filled in each time it is called.Chapter 1: Objects Stored Procedures SQL Reference: Fundamentals 51 Modifying Stored Procedures You can modify a stored procedure definition using the REPLACE PROCEDURE statement. Example Assume you want to change the previous example to insert salary information to the Employee table for new employees. The REPLACE PROCEDURE statement looks like this: REPLACE PROCEDURE NewProc (IN name CHAR(12), IN number INTEGER, IN dept INTEGER, IN salary DECIMAL(10,2), OUT dname CHAR(10), INOUT errstr VARCHAR(30)) BEGIN DECLARE CONTINUE HANDLER FOR SQLSTATE VALUE '23505' SET errstr = 'Duplicate Row.'; INSERT INTO Employee (EmpName, EmpNo, DeptNo, Salary_Amount) VALUES (name, number, dept, salary); SELECT DeptName INTO dname FROM Department WHERE DeptNo = dept; END; Executing Stored Procedures You can execute a stored procedure from any supporting client utility or interface using the SQL CALL statement. You have to specify arguments for all the parameters contained in the stored procedure. The CALL statement for executing the procedure created in the CREATE PROCEDURE example looks like this: CALL NewProc (Jonathan, 1066, 34, dname); For details on executing stored procedures and on call arguments, see “CALL” in SQL Reference: Data Manipulation Statements. Recompiling Stored Procedures The ALTER PROCEDURE feature enables recompilation of stored procedures without having to execute SHOW PROCEDURE and REPLACE PROCEDURE statements. This feature provides the following benefits: • Stored procedures created in earlier releases of Teradata Database can be recompiled in Teradata Database release V2R5.0 and later to derive the benefits of new features and performance improvements. • Recompilation is also useful for cross-platform archive and restoration of stored procedures.Chapter 1: Objects Stored Procedures 52 SQL Reference: Fundamentals • ALTER PROCEDURE allows changes in the following compile-time attributes of a stored procedure: • SPL option • Warnings option Note: For stored procedures created in Teradata Database release V2R5.0 and later to work in earlier releases, they must be recompiled. Deleting Stored Procedures You can delete a stored procedure from a database using the DROP PROCEDURE statement. Assume you want to drop the NewProc procedure from the database. The DROP PROCEDURE statement looks like this: DROP PROCEDURE NewProc; Renaming Stored Procedures You can rename a stored procedure using the RENAME PROCEDURE statement. Assume you want to rename the NewProc stored procedure as NewEmp. The statement looks like this: RENAME PROCEDURE NewProc TO NewEmp; Getting Stored Procedure Information You can get information about the parameters specified in a stored procedure and their attributes using the HELP PROCEDURE statement. The output contains a list of all the parameters specified in the procedure and the attributes of each parameter. The statement to specify is: HELP PROCEDURE NewProc; To view the creation-time attributes of the stored procedure, specify the following statement: HELP PROCEDURE NewProc ATTRIBUTES; Archiving Procedures Stored procedures are archived and restored as part of a database archive and restoration. Individual stored procedures cannot be archived or restored using the ARCHIVE (DUMP) or RESTORE statements. Related Topics FOR details on … SEE … stored procedure control and condition handling statements SQL Reference: Stored Procedures and Embedded SQL invoking stored procedures the CALL statement in SQL Reference: Data Manipulation StatementsChapter 1: Objects External Stored Procedures SQL Reference: Fundamentals 53 External Stored Procedures Introduction External stored procedures are written in the C or C++ programming language, installed on the database, and then executed like stored procedures. Usage Here is a synopsis of the steps you take to develop, compile, install, and use external stored procedures: 1 If you are creating a new external stored procedure, then write, test, and debug the C or C++ code for the procedure. -orIf you are using a third party object or package, then skip to the next step. 2 Use CREATE PROCEDURE or REPLACE PROCEDURE for external stored procedures to identify the location of the source code, object, or package, and install it on the server. The external stored procedure is compiled, if the source code is submitted, linked to the dynamic linked library (DLL or SO) associated with the database in which the procedure resides, and distributed to all Teradata Database nodes in the system. 3 Use GRANT to grant privileges to users who are authorized to use the external stored procedure. 4 Invoke the procedure using the CALL statement. Differences Between Stored Procedures and External Stored Procedures Using external stored procedures is very similar to using stored procedures, except for the following: • Unlike stored procedures, external stored procedures cannot contain any embedded SQL statements. To call a stored procedure, an external stored procedure can call the FNC_CallSP library function. • Invoking an external stored procedure from a client application does not affect the nesting limit for stored procedures. creating or replacing stored procedures SQL Reference: Data Definition Statements dropping stored procedures renaming stored procedures FOR details on … SEE …Chapter 1: Objects User-Defined Functions 54 SQL Reference: Fundamentals • The CREATE PROCEDURE statement for external stored procedures is different from the CREATE PROCEDURE statement for stored procedures. In addition to syntax differences, you do not have to use the COMPILE command in BTEQ or BTEQWIN. • To install an external stored procedure on a database, you must have the CREATE EXTERNAL PROCEDURE privilege on the database. Related Topics User-Defined Functions Introduction SQL provides a set of useful functions, but they might not satisfy all of the particular requirements you have to process your data. User-defined functions (UDFs) allow you to extend SQL by writing your own functions in the C or C++ programming language, installing them on the database, and then using them like standard SQL functions. You can also install UDF objects or packages from third-party vendors, without providing the source code. UDF Types Teradata Database supports three types of UDFs. FOR details on … SEE … external stored procedure programming SQL Reference: UDF, UDM, and External Stored Procedure Programming invoking external stored procedures the CALL statement in SQL Reference: Data Manipulation Statements installing external stored procedures on the server the CREATE/REPLACE PROCEDURE statement in SQL Reference: Data Definition Statements UDF Type Description Scalar Scalar functions take input parameters and return a single value result. Examples of standard SQL scalar functions are CHARACTER_LENGTH, POSITION, and TRIM. Aggregate Aggregate functions produce summary results. They differ from scalar functions in that they take grouped sets of relational data, make a pass over each group, and return one result for the group. Some examples of standard SQL aggregate functions are AVG, SUM, MAX, and MIN. Table A table function is invoked in the FROM clause of a SELECT statement and returns a table to the statement.Chapter 1: Objects Profiles SQL Reference: Fundamentals 55 Usage Here is a synopsis of the steps you take to develop, compile, install, and use a UDF: 1 If you are creating a new UDF, then write, test, and debug the C or C++ code for the UDF. -orIf you are using a third party UDF object or package, then skip to the next step. 2 Use CREATE FUNCTION or REPLACE FUNCTION to identify the location of the source code, object, or package, and install it on the server. The function is compiled, if the source code is submitted, linked to the dynamic linked library (DLL or SO) associated with the database in which the function resides, and distributed to all Teradata Database nodes in the system. 3 Use GRANT to grant privileges to users who are authorized to use the UDF. 4 Call the function. Related Topics Profiles Definition Profiles define values for the following system parameters: • Default database • Spool space • Temporary space • Default account and alternate accounts • Password security attributes An administrator can define a profile and assign it to a group of users who share the same settings. FOR more information on … SEE … writing, testing, and debugging source code for a UDF SQL Reference: UDF, UDM, and External Stored Procedure Programming data definition statements related to UDFs, including CREATE FUNCTION and REPLACE FUNCTION SQL Reference: Data Definition StatementsChapter 1: Objects Profiles 56 SQL Reference: Fundamentals Advantages of Using Profiles Use profiles to: • Simplify system administration. Administrators can create a profile that contains system parameters and assign the profile to a group of users. To change a parameter, the administrator updates the profile instead of each individual user. • Control password security. A profile can define password attributes such as the number of: • Days before a password expires • Days before a password can be used again • Minutes to lock out a user after a certain number of failed logon attempts Administrators can assign the profile to an individual user or to a group of users. Usage The following steps describe how to use profiles to manage a common set of parameters for a group of users. 1 Define a user profile. A CREATE PROFILE statement defines a profile, and lets you set: • Account identifiers to charge for the space used and a default account identifier • Default database • Space to allocate for spool files • Space to allocate for temporary tables • Number of days before the password expires • Minimum and maximum number of characters in a password string • Whether or not to allow digits and special characters in a password string • Number of incorrect logon attempts to allow before locking a user • Number of minutes before unlocking a locked user • Number of days before a password can be used again 2 Assign the profile to users. Use the CREATE USER or MODIFY USER statement to assign a profile to a user. Profile settings override the values set for the user. 3 If necessary, change any of the system parameters for a profile. Use the MODIFY PROFILE statement to change a profile. Related Topics For information on the syntax and usage of profiles, see SQL Reference: Data Definition Statements.Chapter 1: Objects Roles SQL Reference: Fundamentals 57 Roles Definition Roles define access privileges on database objects. A user who is assigned a role can access all the objects that the role has privileges to. Roles simplify management of user access rights. A database administrator can create different roles for different job functions and responsibilities, grant specific privileges on database objects to the roles, and then grant membership to the roles to users. Advantages of Using Roles Use roles to: • Simplify access rights administration. A database administrator can grant rights on database objects to a role and have the rights automatically applied to all users assigned to that role. When a user’s function within an organization changes, changing the user’s role is far easier than deleting old rights and granting new rights to go along with the new function. • Reduce dictionary disk space. Maintaining rights on a role level rather than on an individual level makes the size of the DBC.AccessRights table much smaller. Instead of inserting one row per user per right on a database object, the Teradata Database inserts one row per role per right in DBC.AccessRights, and one row per role member in DBC.RoleGrants. Usage The following steps describe how to manage user access privileges using roles. 1 Define a role. A CREATE ROLE statement defines a role. A newly created role does not have any associated privileges. 2 Add access privileges to the role. Use the GRANT statement to grant privileges to roles on databases, tables, views, macros, columns, triggers, stored procedures, join indexes, hash indexes, and user-defined functions. 3 Grant the role to users or other roles. Use the GRANT statement to grant a role to users or other roles.Chapter 1: Objects User-Defined Types 58 SQL Reference: Fundamentals 4 Assign default roles to users. Use the DEFAULT ROLE option of the CREATE USER or MODIFY USER statement to specify the default role for a user, where: At logon time, the default role of the user becomes the current role for the session. Rights validation uses the active roles for a user, which include the current role and all nested roles. 5 If necessary, change the current role for a session. Use the SET ROLE statement to change the current role for a session. Managing role-based access rights requires sufficient privileges. For example, the CREATE ROLE statement is only authorized to users who have the CREATE ROLE system privilege. Related Topics For information on the syntax and usage of roles, see SQL Reference: Data Definition Statements. User-Defined Types Introduction SQL provides a set of predefined data types, such as INTEGER and VARCHAR, that you can use to store the data that your application uses, but they might not satisfy all of the requirements you have to model your data. User-defined types (UDTs) allow you to extend SQL by creating your own data types and then using them like predefined data types. DEFAULT ROLE = … Specifies … role_name the name of one role to assign as the default role for a user. NONE NULL that the user does not have a default role. ALL the default role to be all roles that are directly or indirectly granted to the user.Chapter 1: Objects User-Defined Types SQL Reference: Fundamentals 59 UDT Types Teradata Database supports distinct and structured UDTs. Distinct and structured UDTs can define methods that operate on the UDT. For example, a distinct UDT named euro can define a method that converts the value to a US dollar amount. Similarly, a structured UDT named circle can define a method that computes the area of the circle using the radius attribute. Using a Distinct UDT Here is a synopsis of the steps you take to develop and use a distinct UDT: 1 Use the CREATE TYPE statement to create a distinct UDT that is based on a predefined data type, such as INTEGER or VARCHAR. The Teradata Database automatically generates functionality for the UDT that allows you to import and export the UDT between the client and server, use the UDT in a table, perform comparison operations between two UDTs, and perform data type conversions between the UDT and the predefined data type on which the definition is based. 2 If the UDT defines methods, write, test, and debug the C or C++ code for the methods, and then use CREATE METHOD or REPLACE METHOD to identify the location of the source code and install it on the server. The methods are compiled, linked to the dynamic linked library (DLL or SO) associated with the SYSUDTLIB database, and distributed to all Teradata Database nodes in the system. 3 Use GRANT to grant privileges to users who are authorized to use the UDT. 4 Use the UDT as the data type of a column in a table definition. UDT Type Description Example Distinct A UDT that is based on a single predefined data type, such as INTEGER or VARCHAR. A distinct UDT named euro that is based on a DECIMAL(8,2) data type can store monetary data. Structured A UDT that is a collection of one or more fields called attributes, each of which is defined as a predefined data type or other UDT (which allows nesting). A structured UDT named circle can consist of x-coordinate, y-coordinate, and radius attributes.Chapter 1: Objects User-Defined Types 60 SQL Reference: Fundamentals Using a Structured UDT Here is a synopsis of the steps you take to develop and use a structured UDT: 1 Use the CREATE TYPE statement to create a structured UDT and specify attributes, constructor methods, and instance methods. Teradata Database automatically generates the following functionality: • A default constructor function that you can use to construct a new instance of the structured UDT and initialize the attributes to NULL • Observer methods for each attribute that you can use to get the attribute values • Mutator methods for each attribute that you can use to set the attribute values 2 Follow these steps to implement, install, and register cast functionality for the UDT (Teradata Database does not automatically generate cast functionality for structured UDTs): a Write, test, and debug C or C++ code that implements cast functionality that allows you to perform data type conversions between the UDT and other data types, including other UDTs. b Identify the location of the source code and install it on the server: The source code is compiled, linked to the dynamic linked library (DLL or SO) associated with the SYSUDTLIB database, and distributed to all Teradata Database nodes in the system. c Use the CREATE CAST or REPLACE CAST statement to register the method or function as a cast routine for the UDT. d Repeat Steps a through c for all methods or functions that provide cast functionality. 3 Follow these steps to implement, install, and register ordering functionality for the UDT (Teradata Database does not automatically generate ordering functionality for structured UDTs): a Write, test, and debug C or C++ code that implements ordering functionality that allows you to perform comparison operations between two UDTs. b Identify the location of the source code and install it on the server: IF you write the source code as a … THEN use one of the following statements … method CREATE METHOD or REPLACE METHOD function CREATE FUNCTION or REPLACE FUNCTION IF you write the source code as a … THEN use one of the following statements … method CREATE METHOD or REPLACE METHOD function CREATE FUNCTION or REPLACE FUNCTIONChapter 1: Objects User-Defined Types SQL Reference: Fundamentals 61 The source code is compiled, linked to the dynamic linked library (DLL or SO) associated with the SYSUDTLIB database, and distributed to all Teradata Database nodes in the system. c Use the CREATE ORDERING or REPLACE ORDERING statement to register the method or function as an ordering routine for the UDT. 4 Follow these steps to implement, install, and register transform functionality for the UDT (Teradata Database does not automatically generate transform functionality for structured UDTs): a Write, test, and debug C or C++ code that implements transform functionality that allows you to import and export the UDT between the client and server. b Identify the location of the source code and install it on the server: The source code is compiled, linked to the dynamic linked library (DLL or SO) associated with the SYSUDTLIB database, and distributed to all Teradata Database nodes in the system. c Repeat Steps a through b. d Use the CREATE TRANSFORM or REPLACE TRANSFORM statement to register the transform routines for the UDT. 5 If the UDT defines constructor methods or instance methods, write, test, and debug the C or C++ code for the methods, and then use CREATE METHOD or REPLACE METHOD to identify the location of the source code and install it on the server. IF the source code implements transform functionality for … THEN … importing the UDT to the server you must write the source code as a UDF and use CREATE FUNCTION or REPLACE FUNCTION to identify the location of the source code and install it on the server. exporting the UDT from the server IF you write the source code as a … THEN use one of the following statements to identify the location of the source code and install it on the server … method CREATE METHOD or REPLACE METHOD function CREATE FUNCTION or REPLACE FUNCTION IF you took Steps a through b to implement and install this transform functionality … THEN repeat Steps a through b to implement and install this transform functionality … importing the UDT to the server exporting the UDT from the server exporting the UDT from the server importing the UDT to the serverChapter 1: Objects User-Defined Types 62 SQL Reference: Fundamentals The methods are compiled, linked to the dynamic linked library (DLL or SO) associated with the SYSUDTLIB database, and distributed to all Teradata Database nodes in the system. 6 Use GRANT to grant privileges to users who are authorized to use the UDT. 7 Use the UDT as the data type of a column in a table definition. Related Topics FOR more information on … SEE … • CREATE TYPE • CREATE METHOD and REPLACE METHOD • CREATE FUNCTION and REPLACE FUNCTION • CREATE CAST and REPLACE CAST • CREATE ORDERING and REPLACE ORDERING • CREATE TRANSFORM and REPLACE TRANSFORM SQL Reference: Data Definition Statements writing, testing, and debugging source code for a constructor method or instance method SQL Reference: UDF, UDM, and External Stored Procedure ProgrammingSQL Reference: Fundamentals 63 CHAPTER 2 Basic SQL Syntax and Lexicon This chapter explains the syntax and lexicon for Teradata SQL, a single, unified, nonprocedural language that provides capabilities for queries, data definition, data modification, and data control of the Teradata Database. Topics include: • Structure of an SQL statement • Keywords • Expressions • Names • Literals • Operators • Functions • Delimiters • Separators • Comments • Terminators • Null statements Structure of an SQL Statement Syntax The following diagram indicates the basic structure of an SQL statement. FF07D232 statement_keyword ; expressions functions keywords clauses , phrasesChapter 2: Basic SQL Syntax and Lexicon Structure of an SQL Statement 64 SQL Reference: Fundamentals where: Typical SQL Statement A typical SQL statement consists of a statement keyword, one or more column names, a database name, a table name, and one or more optional clauses introduced by keywords. For example, in the following single-statement request, the statement keyword is SELECT: SELECT deptno, name, salary FROM personnel.employee WHERE deptno IN(100, 500) ORDER BY deptno, name ; The select list for this statement is made up of the names: • Deptno, name, and salary (the column names) • Personnel (the database name) • Employee (the table name) The search condition, or WHERE clause, is introduced by the keyword WHERE. WHERE deptno IN(100, 500) The sort order, or ORDER BY, clause is introduced by the keywords ORDER BY. ORDER BY deptno, name This syntax element … Specifies … statement_keyword the name of the statement. expressions literals, name references, or operations using names and literals. functions the name of a function and its arguments, if any. keywords special values introducing clauses or phrases or representing special objects, such as NULL. Most keywords are reserved words and cannot be used in names. clauses subordinate statement qualifiers. phrases data attribute phrases. ; the Teradata SQL statement separator and request terminator. The semicolon separates statements in a multistatement request and terminates a request when it is the last non-blank character on an input line in BTEQ. Note that the request terminator is required for a request defined in the body of a macro. For a discussion of macros and their use, see “Macros” on page 46.Chapter 2: Basic SQL Syntax and Lexicon SQL Lexicon Characters SQL Reference: Fundamentals 65 Related Topics The pages that follow provide details on the elements that appear in an SQL statement. SQL Lexicon Characters Client Character Data The characters that make up the SQL lexicon can be represented on the client system in ASCII, EBCDIC, UTF8, UTF16, or in an installed user-defined character set. If the client system character data is not ASCII, then it is converted by the Teradata Database to an internal form for processing and storage. Data returned to the client system is converted to the client character set. Server Character Data The internal forms used for character support are described in International Character Set Support. The notation used for Japanese characters is described in: • “Character Shorthand Notation Used In This Book” • Appendix A: “Notation Conventions.” Case Sensitivity See the following topics in SQL Reference: Data Types and Literals: • “Defining Case Sensitivity for Table Columns” • “CASESPECIFIC Phrase” • “UPPERCASE Phrase” • "Character Data Literals" FOR more information on … SEE … statement_keyword “Keywords” on page 66 keywords expressions “Expressions” on page 67 functions “Functions” on page 92 separators “Separators” on page 94 terminators “Terminators” on page 96Chapter 2: Basic SQL Syntax and Lexicon Keywords 66 SQL Reference: Fundamentals See the following topics in SQL Reference: Functions and Operators: • “LOWER Function” • “UPPER Function” Keywords Introduction Keywords are words that have special meanings in SQL statements. There are two types of keywords: reserved and non-reserved. You cannot use reserved keywords to name database objects. Although you can use non-reserved keywords as object names, you usually should not because of possible confusion resulting from their use. Statement Keyword The statement keyword, the first keyword in an SQL statement, is usually a verb. For example, in the INSERT statement, the first keyword is INSERT. Keywords Other keywords appear throughout a statement as modifiers (for example, DISTINCT, PERMANENT), or as words that introduce clauses (for example, IN, AS, AND, TO, WHERE). In this book, keywords appear entirely in uppercase letters, though SQL does not discriminate between uppercase and lowercase letters in a keyword. For example, SQL interprets the following SELECT statements to be identical: Select Salary from Employee where EmpNo = 10005; SELECT Salary FROM Employee WHERE EmpNo = 10005; select Salary FRom Employee WherE EmpNo = 10005; All keywords must be from the ASCII repertoire. Fullwidth letters are not valid regardless of the character set being used. For a list of Teradata SQL keywords, see Appendix B: “Restricted Words for V2R6.2.” Keywords and Object Names Note that you cannot use reserved keywords to name database objects. Because new keywords are frequently added to new releases of the Teradata Database, you may experience a problem with database object names that were valid in prior releases but which become nonvalid in a new release. The workaround for this is to do one of the following things: • Put the newly nonvalid name in double quotes. • Rename the object. In either case you must change your applications.Chapter 2: Basic SQL Syntax and Lexicon Expressions SQL Reference: Fundamentals 67 Expressions Introduction An expression specifies a value. An expression can consist of literals (or constants), name references, or operations using names and literals. Scalar Expressions A scalar expression, or value expression, produces a single number, character string, byte string, date, time, timestamp, or interval. A value expression has exactly one declared type, common to every possible result of evaluation. Implicit type conversion rules apply to expressions. Query Expressions Query expressions operate on table values and produce rows and tables of data. Every query expression includes at least one FROM clause, which operates on a table reference and returns a single table value. Related Topics Names Introduction In Teradata SQL, various database objects such as tables, views, stored procedures, macros, columns, and collations are identified by a name. The set of valid names depends on whether the system is enabled for Japanese language support. FOR more information on … SEE … • CASE expresssions • arithmetic expressions • logical expressions • datetime expressions • interval expressions • character expresssions • byte expressions SQL Reference: Functions and Operators. data type conversions SQL Reference: Functions and Operators. query expressions SQL Reference: Data Manipulation Statements.Chapter 2: Basic SQL Syntax and Lexicon Names 68 SQL Reference: Fundamentals Rules The rules for naming Teradata Database database objects on systems enabled for standard language support are as follows. • You must define and reference each object, such as user, database, or table, by a name. • In general, names consist of 1 to 30 characters. • Names can appear as a sequence of characters within double quotes and as a quoted hexadecimal string followed by the key letters XN. Such names have fewer restrictions on the characters that can be included. The restrictions are described in “QUOTATION MARKS Characters and Names” on page 69 and “Internal Hexadecimal Representation of a Name” on page 70. • Unquoted names have the following syntactic restrictions: • They may only include the following characters: • Uppercase or lowercase letters (A to Z and a to z) • Digits (0 through 9) • The special characters DOLLAR SIGN ($), NUMBER SIGN (#), and LOW LINE ( _ ) • They must not begin with a digit. • They must not be a keyword. • Systems that are enabled for Japanese language support allow various Japanese characters to be used for names, but determining the maximum number of characters allowed in a name becomes much more complex (see “Name Validation on Systems Enabled with Japanese Language Support” on page 77). • Names having any of the following characteristics are not ANSI SQL-2003 compliant: • Contains lower case letters. • Contains either a $ or a #. • Begins with an underscore. • Has more than 18 characters. • Names that define databases and objects must observe the following rules. • Databases, users, and profiles must have unique names. • Tables, views, stored procedures, join or hash indexes, triggers, user-defined functions, or macros can take the same name as the database or user in which they are created, but cannot take the same name as another of these objects in the same database or user. • Roles can have the same name as a profile, table, column, view, macro, trigger, table function, user-defined function, external stored procedure, or stored procedure; however, role names must be unique among users and databases. • Table and view columns must have unique names. • Parameters defined for a macro or stored procedure must have unique names. • Secondary indexes on a table must have unique names.Chapter 2: Basic SQL Syntax and Lexicon Names SQL Reference: Fundamentals 69 • Named constraints on a table must have unique names. • Secondary indexes and constraints can have the same name as the table they are associated with. • CHECK constraints, REFERENCE constraints, and INDEX objects can also have assigned names. Names are optional for these objects. • Names are not case-specific (see “Case Sensitivity and Names” on page 71). QUOTATION MARKS Characters and Names Enclosing names in QUOTATION MARKS characters (U+0022) greatly increases the valid set of characters for defining names. Pad characters and special characters can also be included. For example, the following strings are both valid names. • “Current Salary” • “D’Augusta” The QUOTATION MARKS characters are not part of the name, but they are required, if the name is not valid otherwise. For example, these two names are identical, even though one is enclosed within QUOTATION MARKS characters. • This_Name • “This_Name” On systems enabled for standard language support, any character translatable to the LATIN server character set can appear in an object name, with the following exceptions: • The NULL character (U+0000) is not allowed in any names, including quoted names. • The object name must not consist entirely of blank characters. In this context, a blank character is any of the following: • NULL (U+0000) • CHARACTER TABULATION (U+0009) • LINE FEED (U+000A) • LINE TABULATION (U+000B) • FORM FEED (U+000C) • CARRIAGE RETURN (U+000D) • SPACE (U+0020) • The code point 0x1A, which represents the error character for KANJI1 and LATIN server character sets, cannot be translated between character sets and must not appear in object names. All of the following examples are valid names. • Employee • job_title Chapter 2: Basic SQL Syntax and Lexicon Names 70 SQL Reference: Fundamentals • CURRENT_SALARY • DeptNo • Population_of_Los_Angeles • Totaldollars • “Table A” • “Today’s Date” Note: If you use quoted names, the QUOTATION MARKS characters that delineate the names are not counted in the length of the name and are not stored in Dictionary tables used to track name usage. If a Dictionary view is used to display such names, they are displayed without the double quote characters, and if the resulting names are used without adding double quotes, the likely outcome is an error report. For example, “D’Augusta” might be the name of a column in the Dictionary view DBC.Columns, and the HELP statements that return column names return the name as D’Augusta (without being enclosed in QUOTATION MARKS characters). Internal Hexadecimal Representation of a Name You can also create and reference object names by their internal hexadecimal representation in the Data Dictionary using the following syntax: where: The key letters XN specify that the string is a hexadecimal name. On systems enabled for standard language support, any character translatable to the LATIN server character set can appear in an object name, with the same exceptions listed in the preceding section, “QUOTATION MARKS Characters and Names” on page 69. For more information on using internal hexadecimal representations of names, see “Using the Internal Hexadecimal Representation of a Name” on page 82. Syntax element … Specifies … 'hexadecimal_digits' a quoted hexadecimal string representation of the Teradata Database internal encoding. HH01A099 'hexadecimal_digit(s)' XNChapter 2: Basic SQL Syntax and Lexicon Standard Form for Data in Teradata Database SQL Reference: Fundamentals 71 Case Sensitivity and Names Names are not case-dependent—a name cannot be used twice by changing its case. Any mix of uppercase and lowercase can be used when referencing symbolic names in a request. For example, the following statements are identical. SELECT Salary FROM Employee WHERE EmpNo = 10005; SELECT SALARY FROM EMPLOYEE WHERE EMPNO = 10005; SELECT salary FROM employee WHERE eMpNo = 10005; The case in which a column name is defined can be important. The column name is the default title of an output column, and symbolic names are returned in the same case in which they were defined. For example, assume that the columns in the SalesReps table are defined as follows: CREATE TABLE SalesReps ( last_name VARCHAR(20) NOT NULL, first_name VARCHAR(12) NOT NULL, ... In response to a query that does not define a TITLE phrase, such as the following example, the column names are returned exactly as defined they were defined, for example, last_name, then first_name. SELECT Last_Name, First_Name FROM SalesReps ORDER BY Last_Name; You can use the TITLE phrase to specify the case, wording, and placement of an output column heading either in the column definition or in an SQL statement. For more information, see SQL Reference: Data Manipulation Statements. Standard Form for Data in Teradata Database Introduction Data in Teradata Database is presented to a user according to the relational model, which models data as two dimensional tables with rows and columns. Each row of a table is composed one or more columns identified by column name. Each column contains a data item (or a null) having a single data type. Syntax for Referencing a Column table_name. FF07D238 column_name database_name.Chapter 2: Basic SQL Syntax and Lexicon Standard Form for Data in Teradata Database 72 SQL Reference: Fundamentals where: Definition: Fully Qualified Column Name A fully qualified name consists of a database name, table name, and column name. For example, a fully qualified reference for the Name column in the Employee table of the Personnel database is: Personnel.Employee.Name Column Alias In addition to referring to a column by name, an SQL query can reference a column by an alias. Column aliases are used for join indexes when two columns have the same name. However, an alias can be used for any column when a pseudonym is more descriptive or easier to use. Using an alias to name an expression allows a query to reference the expression. You can specify a column alias with or without the keyword AS on the first reference to the column in the query. The following example creates and uses aliases for the first two columns. SELECT departnumber AS d, employeename e, salary FROM personnel.employee WHERE d IN(100, 500) ORDER BY d, e ; Alias names must meet the same requirements as names of other database objects. For details, see “Names” on page 67. The scope of alias names is confined to the query. Syntax element … Specifies … database_name a qualifying name for the database in which the table and column being referenced is stored. Depending on the ambiguity of the reference, database_name might or might not be required. See “Unqualified Object Names” on page 73. table_name a qualifying name for the table in which the column being referenced is stored. Depending on the ambiguity of the reference, table_name might or might not be required. See “Unqualified Object Names” on page 73. column_name one of the following: • The name of the column being referenced • The alias of the column being referenced • The keyword PARTITION See “Column Alias” on page 72.Chapter 2: Basic SQL Syntax and Lexicon Unqualified Object Names SQL Reference: Fundamentals 73 Referencing All Columns in a Table An asterisk references all columns in a row simultaneously, for example, the following SELECT statement references all columns in the Employee table. A list of those fully qualified column names follows the query. SELECT * FROM Employee; Personnel.Employee.EmpNo Personnel.Employee.Name Personnel.Employee.DeptNo Personnel.Employee.JobTitle Personnel.Employee.Salary Personnel.Employee.YrsExp Personnel.Employee.DOB Personnel.Employee.Sex Personnel.Employee.Race Personnel.Employee.MStat Personnel.Employee.EdLev Personnel.Employee.HCap Unqualified Object Names Definition An unqualified object name is a table, column, trigger, macro, or stored procedure reference that is not fully qualified. For example, the WHERE clause in the following statement uses “DeptNo” as an unqualified column name: SELECT * FROM Personnel.Employee WHERE DeptNo = 100 ; Unqualified Column Names You can omit database and table name qualifiers when you reference columns as long as the reference is not ambiguous. For example, the WHERE clause in the following statement: SELECT Name, DeptNo, JobTitle FROM Personnel.Employee WHERE Personnel.Employee.DeptNo = 100 ; can be written as: WHERE DeptNo = 100 ; because the database name and table name can be derived from the Personnel.Employee reference in the FROM clause.Chapter 2: Basic SQL Syntax and Lexicon Unqualified Object Names 74 SQL Reference: Fundamentals Omitting Database Names When you omit the database name qualifier, Teradata Database looks in the following databases to find the unqualified table, view, trigger, or macro name: • The default database, which is established by a DATABASE, CREATE USER, MODIFY USER, CREATE PROFILE, or MODIFY PROFILE statement • Other databases, if any, referenced by the SQL statement • The login user database for a volatile table, if the unqualified object name is a table name The search must find the table name in only one of those databases. An ambiguous name error message results if the name exists in more than one of those databases. For example, if your login user database has no volatile tables named Employee and you have established Personnel as your default database, you can omit the Personnel database name qualifier from the preceding sample query. Rules for Name Resolution The following rules govern name resolution: • Name resolution is performed statement by statement. • When an INSERT statement contains a subquery, names are resolved in the subquery first. • Names in a view are resolved when the view is created. • Names in a macro data manipulation statement are resolved when the macro is created. • Names in a macro data definition statement are resolved when the macro is performed using the default database of the user submitting the EXECUTE statement. Therefore, you should fully qualify all names in a macro data definition statement, unless you specifically intend for the user’s default to be used. • Names in stored procedure statements are resolved when the procedure is created. All unqualified object names acquire the current default database name. • An ambiguous unqualified name returns an error to the requestor. Related Topics FOR more information on … SEE … default databases “Default Database” on page 75. the DATABASE statement “SQL Data Definition Language Statement Syntax” in SQL Reference: Data Definition Statements. the CREATE USER statement the MODIFY USER statementChapter 2: Basic SQL Syntax and Lexicon Default Database SQL Reference: Fundamentals 75 Default Database Definition The default database is a Teradata extension to SQL that defines a database that Teradata Database uses to look for unqualified table, view, trigger, or macro names in SQL statements. The default database is not the only database that Teradata Database uses to find an unqualified table, view, trigger, or macro name in an SQL statement, however; Teradata Database also looks for the name in: • Other databases, if any, referenced by the SQL statement • The login user database for a volatile table, if the unqualified object name is a table name If the unqualified object name exists in more than one of the databases in which Teradata Database looks, the SQL statement produces an ambiguous name error. Establishing a Permanent Default Database You can establish a permanent default database that is invoked each time you log on. For example, the following statement automatically establishes Personnel as the default database for Marks at the next logon: MODIFY USER marks AS DEFAULT DATABASE = personnel ; After you assign a default database, Teradata Database uses that database as one of the databases to look for all unqualified object references. To obtain information from a table, view, trigger, or macro in another database, fully qualify the table reference by specifying the database name, a FULLSTOP character, and the table name. TO … USE one of the following SQL Data Definition statements … define a permanent default database • CREATE USER, with a DEFAULT DATABASE clause. • CREATE USER, with a PROFILE clause that specifies a profile that defines the default database. change your permanent default database definition • MODIFY USER, with a DEFAULT DATABASE clause. • MODIFY USER, with a PROFILE clause. • MODIFY PROFILE, with a DEFAULT DATABASE clause. add a default database when one had not been established previouslyChapter 2: Basic SQL Syntax and Lexicon Default Database 76 SQL Reference: Fundamentals Establishing a Default Database for a Session You can establish a default database for the current session that Teradata Database uses to look for unqualified table, view, trigger, or macro names in SQL statements. For example, after entering the following SQL statement: DATABASE personnel ; you can enter a SELECT statement as follows: SELECT deptno (TITLE 'Org'), name FROM employee ; which has the same results as: SELECT deptno (TITLE 'Org'), name FROM personnel.employee; To establish a default database, you must have some privilege on a database, macro, stored procedure, table, user, or view in that database. Once defined, the default database remains in effect until the end of a session or until it is replaced by a subsequent DATABASE statement. Related Topics TO … USE … establish a default database for a session the DATABASE statement. FOR more information on … SEE … the DATABASE statement SQL Reference: Data Definition Statements. the CREATE USER statement the MODIFY USER statement fully-qualified names “Standard Form for Data in Teradata Database” on page 71. “Unqualified Object Names” on page 73. using profiles to define a default database “Profiles” on page 55.Chapter 2: Basic SQL Syntax and Lexicon Name Validation on Systems Enabled with Japanese Language Support SQL Reference: Fundamentals 77 Name Validation on Systems Enabled with Japanese Language Support Introduction A system that is enabled with Japanese language support allows thousands of additional characters to be used for names, but also introduces additional restrictions. Rules for Unquoted Names Unquoted names can use the following characters when Japanese language support is enabled: • Any character valid in an unquoted name under standard language support: • Uppercase or lowercase letters (A to Z and a to z) • Digits (0 through 9) • The special characters DOLLAR SIGN ($), NUMBER SIGN (#), and LOW LINE ( _ ) • The fullwidth (zenkaku) versions of the characters valid for names under standard language support: • Fullwidth uppercase or lowercase letters (A to Z and a to z) • Fullwidth digits (0 through 9) • The special characters fullwidth DOLLAR SIGN ($), fullwidth NUMBER SIGN (#), and fullwidth LOW LINE ( _ ) • Fullwidth (zenkaku) and halfwidth (hankaku) Katakana characters and sound marks. • Hiragana characters. • Kanji characters from JIS-x0208. The length of a name is restricted in a complex fashion. Charts of the supported Japanese character sets, the Teradata Database internal encodings, the valid character ranges for Japanese object names and data, and the non-valid character ranges for Japanese data and object names are documented in International Character Set Support. Rules for Quoted Names and Internal Hexadecimal Representation of Names As described in “QUOTATION MARKS Characters and Names” on page 69 and “Internal Hexadecimal Representation of a Name” on page 70, names can also appear as a sequence of characters within double quotes or as a quoted hexadecimal string followed by the key letters XN. Such names have fewer restrictions on the characters that can be included. The following restrictions that apply to systems enabled for standard language support also apply to systems enabled for Japanese language support: • The NULL character (U+0000) is not allowed. • The code point 0x1A, which represents the error character for KANJI1 and LATIN server character sets, cannot be translated between character sets and must not appear in object names.Chapter 2: Basic SQL Syntax and Lexicon Name Validation on Systems Enabled with Japanese Language Support 78 SQL Reference: Fundamentals • The object name must not consist entirely of blank characters. In this context, a blank character is any of the following: • NULL (U+0000) • LINE FEED (U+000A) • LINE TABULATION (U+000B) • FORM FEED (U+000C) • CARRIAGE RETURN (U+000D) • SPACE (U+0020) • CHARACTER TABULATION (U+0009) Additional rules apply to sessions using non-Japanese client character sets on systems enabled with Japanese language support. Here are some examples of predefined non-Japanese client character sets (you can also define your own site-defined client character sets): For sessions using non-Japanese client character sets on systems where Japanese language support is enabled, object names can only have characters in the following inclusive ranges: • U+0001 through U+000D • U+0015 through U+005B • U+005D through U+007D • U+007F REVERSE SOLIDUS (U+005C) and TILDE (U+007E) are not allowed. Cross-Platform Integrity If you need to access objects from heterogeneous clients, the best practice is to restrict the object names to those allowed under standard language support. Calculating the Length of a Name The length of a name is measured by the physical bytes of its internal representation, not by the number of viewable characters. Under the KanjiEBCDIC character sets, the Shift-Out and Shift-In characters that delimit a multibyte character string are included in the byte count. • EBCDIC • EBCDIC037_0E • ASCII • LATIN1_0A • LATIN9_0A • LATIN1252_0A • UTF8 • UTF16 • SCHEBCDIC935_2IJ • TCHEBCDIC937_3IB • HANGULEBCDIC933_1II • SCHGB2312_1T0 • TCHBIG5_1R0 • HANGULKSC5601_2R4 • SCHGB2312_1T0 • TCHBIG5_1R0 • HANGULKSC5601_2R4Chapter 2: Basic SQL Syntax and Lexicon Name Validation on Systems Enabled with Japanese Language Support SQL Reference: Fundamentals 79 For example, the following table name contains six logical characters of mixed single byte characters/multibyte characters, defined during a KanjiEBCDIC session: QR All single byte characters, including the Shift-Out/Shift-In characters, are translated into the Teradata Database internal encoding, based on JIS-x0201. Under the KanjiEBCDIC character sets, all multibyte characters remain in the client encoding. Thus, the processed name is stored as a string of twelve bytes, padded on the right with the single byte space character to a total of 30 bytes. The internal representation is as follows: 0E 42E3 42C1 42C2 42F1 0F 51 52 20 20 20 20 20 20 20 20 20 20 20 20 ... < T A B 1 > Q R To ensure upgrade compatibility, an object name created under one character set cannot exceed 30 bytes in any supported character set. For example, a single Katakana character occupies 1 byte in KanjiShift-JIS. However, when KanjiShift-JIS is converted to KanjiEUC, each Katakana character occupies two bytes. Thus, a 30-byte Katakana name in KanjiShift-JIS would expand in KanjiEUC to 60 bytes, which is illegal. The formula for calculating the correct length of an object name is as follows: Length = ASCII + (2*KANJI) + MAX (2*KATAKANA, (KATAKANA + 2*S2M + 2*M2S)) where: How Validation Occurs Name validation occurs when the object is created or renamed, as follows: • User names, database names, and account names are verified during the CREATE/ MODIFY USER and CREATE/MODIFY DATABASE statements. • Names of work tables and error tables are validated by the MultiLoad and FastLoad client utilities. • Table names and column names are verified during the CREATE/ALTER TABLE and RENAME TABLE statements. View and macro names are verified during the CREATE/ RENAME VIEW and CREATE/RENAME MACRO statements. This variable … Represents the number of … ASCII single-byte ASCII characters in the name. KATAKANA single-byte Hankaku Katakana characters in the name. KANJI double-byte characters in the name from the JIS-x0208 standard. S2M transitions from ASCII or KATAKANA to JIS-x0208. M2S transitions from JIS-x0208 to ASCII or KATAKANA.Chapter 2: Basic SQL Syntax and Lexicon Name Validation on Systems Enabled with Japanese Language Support 80 SQL Reference: Fundamentals Stored procedure names are verified during the execution of CREATE/RENAME/ REPLACE PROCEDURE statements. • Alias object names used in the SELECT, UPDATE, and DELETE statements are verified. The validation occurs only when the SELECT statement is used in a CREATE/REPLACE VIEW statement, and when the SELECT, UPDATE, or DELETE TABLE statement is used in a CREATE/REPLACE MACRO statement. Examples of Validating Japanese Object Names The following tables illustrate valid and non-valid object names under the Japanese character sets: KanjiEBCDIC, KanjiEUC, and KanjiShift-JIS. The meanings of ASCII, KATAKANA, KANJI, S2M, M2S, and LEN are defined in “Calculating the Length of a Name” on page 78. KanjiEBCDIC Object Name Examples KanjiEUC Object Name Examples Name ASCII Katakana Kanji S2M M2S LEN Result 0 0 14 1 1 32 Not valid because LEN > 30. kl 2 0 12 2 2 34 Not valid because LEN > 30. kl<> 2 0 10 2 2 30 Not valid because consecutive SO and SI characters are not allowed. 0 0 11 2 2 30 Not valid because consecutive SI and SO characters are not allowed. ABCDEFGHIJKLMNO 0 15 0 0 0 30 Valid. KLMNO 0 5 10 1 1 30 Valid. 0 0 1 1 1 6 Not valid because the double byte space is not allowed. Name ASCII Katakana Kanji S2M M2S LEN Result ABCDEFGHIJKLM 6 0 7 3 3 32 Not valid because LEN > 30 bytes. ABCDEFGHIJKLM 6 0 7 2 2 28 Valid. ss 2ABCDEFGHIJKL 0 1 11 1 1 27 Valid. Ass 2 BCDEFGHIJKL 0 1 11 2 2 31 Not valid because LEN > 30 bytes. ss 3C 0 0 0 1 1 4 Not valid because characters from code set 3 are not allowed.Chapter 2: Basic SQL Syntax and Lexicon Object Name Translation and Storage SQL Reference: Fundamentals 81 KanjiShift-JIS Object Name Examples Related Topics For charts of the supported Japanese character sets, the Teradata Database internal encodings, the valid character ranges for Japanese object names and data, and the non-valid character ranges for Japanese data and object names, see International Character Set Support. Object Name Translation and Storage Object names are stored in the dictionary tables using the following translation conventions. Both the ASCII character set and the EBCDIC character set are stored on the server as ASCII. Name ASCII Katakana Kanji S2M M2S LEN Result ABCDEFGHIJKLMNOPQR 6 7 5 1 1 30 Valid. ABCDEFGHIJKLMNOPQR 6 7 5 2 2 31 Not valid because LEN > 30 bytes. Character Type Description Single byte All single byte characters in a name, including the KanjiEBCDIC Shift-Out/ShiftIn characters, are translated into the Teradata Database internal representation (based on JIS-x0201 encoding). Multibyte Multibyte characters in object names are handled according to the character set in effect for the current session, as follows. Multibyte Character Set Description KanjiEBCDIC Each multibyte character within the Shift-Out/ShiftIn delimiters is stored without translation; that is, it remains in the client encoding. The name string must have matched (but not consecutive) Shift-Out and Shift-In delimiters. KanjiEUC Under code set 1, each multibyte character is translated from KanjiEUC to KanjiShift-JIS. Under code set 2, byte ss2 (0x8E) is translated to 0x80; the second byte is left unmodified. This translation preserves the relative ordering of code set 0, code set 1, and code set 2. KanjiShift-JIS Each multibyte character is stored without translation; it remains in the client encoding.Chapter 2: Basic SQL Syntax and Lexicon Object Name Comparisons 82 SQL Reference: Fundamentals Object Name Comparisons Comparison Rules In comparing two names, the following rules apply: • A simple Latin lowercase letter is equivalent to its corresponding simple Latin uppercase letter. For example, 'a' is equivalent to 'A'. • Multibyte characters that have the same logical presentation but have different physical encodings under different character sets do not compare as equivalent. • Two names compare as identical when their internal hexadecimal representations are the same, even if their logical meanings are different under the originating character sets. Note that identical characters on keyboards connected to different clients are not necessarily identical in their internal encoding in the Teradata Database. The Teradata Database could interpret two logically identical names as different names if the character sets under which they were created are not the same. For example, the following strings illustrate the internal representation of two names, both of which were defined with the same logical multibyte characters. However, the first name was created under KanjiEBCDIC, and the second name was created under KanjiShift-JIS. KanjiEBCDIC: 0E 42E3 42C1 42C2 42F1 0F 51 52 KanjiShift-JIS: 8273 8260 8261 8250 D8 D9 To ensure upgrade compatibility, you must avoid semantically duplicate object names in situations where duplicate object names would not normally be allowed. Also, two different character sets might have the same internal encoding for two logically different multibyte characters. Thus, two names might compare as identical even if their logical meanings are different. Using the Internal Hexadecimal Representation of a Name The Teradata Database knows an object name by its internal hexadecimal representation, and this is how it is stored in the various system tables of the Data Dictionary. The encoding of the internal representation of an object name depends on the components of the name string (are there single byte characters, multibyte characters, or both; are there Shift Out/Shift In (SO/SI) characters, and so on) and the character set in effect when the name was created. Suppose that a user under one character set needs to reference an object created by a user under a different character set. If the current user attempts to reference the name with the actual characters (that is, by typing the characters or by selecting non-specific entries from a dictionary table), the access could fail or the returned name could be meaningless. For example, assume that User_1 invokes a session under KanjiEBCDIC and creates a table name with multibyte characters. User_2 invokes a session under KanjiEUC and issues the following statement.Chapter 2: Basic SQL Syntax and Lexicon Object Name Comparisons SQL Reference: Fundamentals 83 SELECT TableName FROM DBC.Tables The result returns the KanjiEBCDIC characters in KanjiEUC presentation, which probably does not make sense. You can avoid this problem by creating objects and specifying object names in the following ways: • Create objects using names that contain only simple single byte Latin letters (A...Z, a...z) digits, and the DOLLAR SIGN ($), NUMBER SIGN (#), and LOW LINE ( _ ) symbols. Because these characters always translate to the same internal representation, they display exactly the same presentation to any session, regardless of the client or the character set. • Use the following syntax to reference a name by its internal representation. where: The key letters XN specify that the string is a hexadecimal name. Example The following table name, which contains mixed single byte characters and multibyte characters, was created under a KanjiEBCDIC character set: KAN The client encoding in which this name was received is as follows: 0E 42E3 42C1 42C2 42F1 0F D2 C1 D5 < T A B 1 > K A N The single byte characters (the letters K, A, and N, and the SO/SI characters) were translated into internal JIS-x0201 encoding. The multibyte characters were not translated and remained in the host encoding. The resulting internal string by which the name was stored is as follows: 0E 42E3 42C1 42C2 42F1 0F 4B 41 4E < T A B 1 > K A N To access this table from a KanjiShift-JIS or KanjiEUC character set, you could use the following Teradata SQL statement: SELECT * FROM '0E42E342C142C242F10F4B414E'XN; The response would be all rows from table KAN. Syntax element … Specifies … ’hexadecimal_digits’ a quoted hexadecimal string representation of the Teradata Database internal encoding. HH01A099 'hexadecimal_digit(s)' XNChapter 2: Basic SQL Syntax and Lexicon Finding the Internal Hexadecimal Representation for Object Names 84 SQL Reference: Fundamentals Finding the Internal Hexadecimal Representation for Object Names Introduction The CHAR2HEXINT function converts a character string to its internal hexadecimal representation. You can use this function to find the internal representation of any Teradata Database name. For more information on CHAR2HEXINT, see SQL Reference: Functions and Operators. Example 1 For example, to find the internal representation of all Teradata Database table names, issue the following Teradata SQL statement. SELECT CHAR2HEXINT(T.TableName) (TITLE 'Internal Hex Representation of TableName'),T.TableName (TITLE 'TableName') FROM DBC.Tables T WHERE T.TableKind = 'T' ORDER BY T.TableName; This statement selects all rows from the DBC.Tables view where the value of the TableKind column is T. For each row selected, both the internal hexadecimal representation and the character format of the value in the TableName column are returned, sorted alphabetically. An example of a portion of the output from this statement is shown below. In this example, the first name (double byte-A) was created using the KanjiEBCDIC character set. Internal Hex Representation of TableName TableName ------------------------------------------------------------ ----------- 0E42C10F2020202020202020202020202020202020202020202020202020 416363657373526967687473202020202020202020202020202020202020 AccessRights 4163634C6F6752756C6554626C2020202020202020202020202020202020 AccLogRuleTb 4163634C6F6754626C202020202020202020202020202020202020202020 AccLogTbl 4163636F756E747320202020202020202020202020202020202020202020 Accounts 416363746720202020202020202020202020202020202020202020202020 Acctg 416C6C202020202020202020202020202020202020202020202020202020 All 4368616E676564526F774A6F75726E616C20202020202020202020202020 ChangedRowJo 636865636B5F7065726D2020202020202020202020202020202020202020 check_perm 436F70496E666F54626C2020202020202020202020202020202020202020 CopInfoTbl Note that the first name, , cannot be interpreted. To obtain a printable version of a name, you must log onto a session under the same character set under which the name was created.Chapter 2: Basic SQL Syntax and Lexicon Finding the Internal Hexadecimal Representation for Object Names SQL Reference: Fundamentals 85 Example 2 You can use the same syntax to obtain the internal hexadecimal representations of all views or all macros. To do this, modify the WHERE condition to TableKind=’V’ for views and to TableKind=’M’ for macros. To obtain the internal hexadecimal representation of all database names, you can issue the following statement: SELECT CHAR2HEXINT(D.DatabaseName)(TITLE 'Internal Hex Representation of DatabaseName'),D.DatabaseName (TITLE 'DatabaseName') FROM DBC.Databases D ORDER BY D.DatabaseName; This statement selects every DatabaseName from DBC.Databases. For each DatabaseName, it returns the internal hexadecimal representation and the name in character format, sorted by DatabaseName. An example of the output from this statement is as follows: Internal Hex Representation of DatabaseName DatabaseName ------------------------------------------------------------ ------------ 416C6C202020202020202020202020202020202020202020202020202020 All 434F4E534F4C452020202020202020202020202020202020202020202020 CONSOLE 437261736864756D70732020202020202020202020202020202020202020 Crashdumps 444243202020202020202020202020202020202020202020202020202020 DBC 44656661756C742020202020202020202020202020202020202020202020 Default 5055424C4943202020202020202020202020202020202020202020202020 PUBLIC 53797341646D696E20202020202020202020202020202020202020202020 SysAdmin 53797374656D466520202020202020202020202020202020202020202020 SystemFe Example 3 Note that these statements return the padded hexadecimal name. The value 0x20 represents a space character in the internal representation. You can use the TRIM function to obtain the hexadecimal values without the trailing spaces, as follows. SELECT CHAR2HEXINT(TRIM(T.TableName)) (TITLE 'Internal Hex Representation of TableName'),T.TableName (TITLE 'TableName') FROM DBC.Tables T WHERE T.TableKind = 'T' ORDER BY T.TableName;Chapter 2: Basic SQL Syntax and Lexicon Specifying Names in a Logon String 86 SQL Reference: Fundamentals Specifying Names in a Logon String Purpose Identifies a user to the Teradata Database and, optionally, permits the user to specify a particular account to log onto. Syntax where: The Teradata Database does not support the hexadecimal representation of a username, a password, or an accountname in a logon string. For example, if you attempt to log on as user DBC by entering '444243'XN, the logon is not successful and an error message is generated. Passwords The password format options allows the site administrator to change the minimum and maximum number of characters allowed in the password string, and control the use of digits and special characters. Password string rules are identical to those for naming objects. See “Name Validation on Systems Enabled with Japanese Language Support” on page 77. The password formatting feature does not apply to multibyte client character sets on systems enabled with Japanese language support. Syntax element … Specifies … tdp_id/username the client TDP the user wishes to use to communicate with the Teradata Database and the name by which the Teradata Database knows the user. The username parameter can contain mixed single byte and multibyte characters if the current character set permits them. password an optional (depending on how the user is defined) password required to gain access to the Teradata Database. The password parameter can contain mixed single byte and multibyte characters if the current character set permits them. accountname an optional account name or account string that specifies a user account or account and performance-related variable parameters the user can use to tailor the session being logged onto. The accountname parameter can contain mixed single byte and multibyte characters if the current character set permits them. tdpid/username HH01A079 ,password ,accountnameChapter 2: Basic SQL Syntax and Lexicon Literals SQL Reference: Fundamentals 87 Literals Literals, or constants, are values coded directly in the text of an SQL statement, view or macro definition text, or CHECK constraint definition text. In general, the system is able to determine the data type of a literal by its form. Numeric Literals A numeric literal (also referred to as a constant) is a character string of 1 to 40 characters selected from the following: • digits 0 through 9 • plus sign • minus sign • decimal point There are three types of numeric literals: integer, decimal, and floating point. Hexadecimal Literals A hexadecimal literal specifies a string of 0 to 62000 hexadecimal digits that can represent a byte, character, or integer value. A hexadecimal digit is a character from 0 to 9, a to f, or A to F. Type Description Integer Literal An integer literal declares literal strings of integer numbers. Integer literals consist of an optional sign followed by a sequence of up to 10 digits. A numeric literal that is outside the range of values of an integer literal is considered a decimal literal. Decimal Literal A decimal literal declares literal strings of decimal numbers. Decimal literals consist of the following components, reading from left-to-right: an optional sign, an optional sequence of up to 38 digits (mandatory only when no digits appear after the decimal point), an optional decimal point, an optional sequence of digits (mandatory only when no digits appear before the decimal point). The scale and precision of a decimal literal are determined by the total number of digits in the literal and the number of digits to the right of the decimal point, respectively. Floating Point Literal A floating point literal declares literal strings of floating point numbers. Floating point literals consist of the following components, reading from left-toright: an optional sign, an optional sequence of digits (mandatory only when no digits appear after the decimal point) representing the whole number portion of the mantissa, an optional decimal point, an optional sequence of digits (mandatory only when no digits appear before the decimal point) representing the fractional portion of the mantissa, the literal character E, an optional sign, a sequence of digits representing the exponent. Chapter 2: Basic SQL Syntax and Lexicon Literals 88 SQL Reference: Fundamentals DateTime Literals Date and time literals declare date, time, or timestamp values in a SQL expression, view or macro definition text, or CONSTRAINT definition text. Date and time literals are introduced by keywords. For example: DATE '1969-12-23' There are three types of DateTime literals: DATE, TIME, and TIMESTAMP. Interval Literals Interval literals provide a means for declaring spans of time. Interval literals are introduced and followed by keywords. For example: INTERVAL '200' HOUR There are two mutually exclusive categories of interval literals: Year-Month and Day-Time. Type Description DATE Literal A date literal declares a date value in ANSI DATE format. ANSI DATE literal is the preferred format for DATE constants. All DATE operations accept this format. TIME Literal A time literal declares a time value and an optional time zone offset. TIMESTAMP Literal A timestamp literal declares a timestamp value and an optional time zone offset. Category Type Description Year-Month • YEAR • YEAR TO MONTH • MONTH Represent a time span that can include a number of years and months. Day-Time • DAY • DAY TO HOUR • DAY TO MINUTE • DAY TO SECOND • HOUR • HOUR TO MINUTE • HOUR TO SECOND • MINUTE • MINUTE TO SECOND • SECOND Represent a time span that can include a number of days, hours, minutes, or seconds.Chapter 2: Basic SQL Syntax and Lexicon Literals SQL Reference: Fundamentals 89 Character Literals A character literal declares a character value in an expression, view or macro definition text, or CHECK constraint definition text. Character literals consist of 0 to 31000 bytes delimited by a matching pair of single quotes. A zero-length character literal is represented by two consecutive single quotes (''). Graphic Literals A graphic literal specifies multibyte characters within the graphic repertoire. Built-In Functions The built-in functions, or special register functions, which are niladic (have no arguments), return various information about the system and can be used like other literals within SQL expressions. In an SQL query, the appropriate system value is substituted by the Parser after optimization but prior to executing a query using a cachable plan. Available built-in functions include all of the following: • ACCOUNT • CURRENT_DATE • CURRENT_TIME • CURRENT_TIMESTAMP • DATABASE • DATE • PROFILE • ROLE • SESSION • TIME • USER Related Topics FOR more information on … SEE … • numeric literals • DateTime literals • interval literals • character literals • graphic literals • hexadecimal literals SQL Reference: Data Types and Literals. built-in functions SQL Reference: Functions and Operators.Chapter 2: Basic SQL Syntax and Lexicon NULL Keyword as a Literal 90 SQL Reference: Fundamentals NULL Keyword as a Literal Null A null represents any of three things: • An empty column • An unknown value • An unknowable value Nulls are neither values nor do they signify values; they represent the absence of value. A null is a place holder indicating that no value is present. NULL Keyword The keyword NULL represents null, and is sometimes available as a special construct similar to, but not identical with, a literal. ANSI Compliance NULL is ANSI SQL-2003-compliant with extensions. Using NULL as a Literal Use NULL as a literal in the following ways: • A CAST source operand, for example: SELECT CAST (NULL AS DATE); • A CASE result, for example. SELECT CASE WHEN orders = 10 THEN NULL END FROM sales_tbl; • An insert item specifying a null is to be placed in a column position on INSERT. • An update item specifying a null is to be placed in a column position on UPDATE. • A default column definition specification, for example: CREATE TABLE European_Sales (Region INTEGER DEFAULT 99 ,Sales Euro_Type DEFAULT NULL); • An explicit SELECT item, for example: SELECT NULL This is a Teradata extension to ANSI. • An operand of a function, for example: SELECT TYPE(NULL) This is a Teradata extension to ANSI. Data Type of NULL When you use NULL as an explicit SELECT item or as the operand of a function, its data type is INTEGER. In all other cases NULL has no data type because it has no value.Chapter 2: Basic SQL Syntax and Lexicon Operators SQL Reference: Fundamentals 91 For example, if you perform SELECT TYPE(NULL), then INTEGER is returned as the data type of NULL. To avoid type issues, cast NULL to the desired type. Related Topics For information on the behavior of nulls and how to use them in data manipulation statements, see “Manipulating Nulls” on page 134. Operators Introduction SQL operators are used to express logical and arithmetic operations. Operators of the same precedence are evaluated from left to right. See “SQL Operations and Precedence” on page 91 for more detailed information. Parentheses can be used to control the order of precedence. When parentheses are present, operations are performed from the innermost set of parentheses outward. Definitions The following definitions apply to SQL operators. SQL Operations and Precedence SQL operations, and the order in which they are performed when no parentheses are present, appear in the following table. Operators of the same precedence are evaluated from left to right. Term Definition numeric Any literal, data reference, or expression having a numeric value. string Any character string or string expression. logical A Boolean expression (resolves to TRUE, FALSE, or unknown). value Any numeric, character, or byte data item. set A collection of values returned by a subquery, or a list of values separated by commas and enclosed by parentheses. Precedence Result Type Operation highest numeric + numeric (unary plus) - numeric (unary minus)Chapter 2: Basic SQL Syntax and Lexicon Functions 92 SQL Reference: Fundamentals Functions Scalar Functions Scalar functions take input parameters and return a single value result. Some examples of standard SQL scalar functions are CHARACTER_LENGTH, POSITION, and SUBSTRING. Aggregate Functions Aggregate functions produce summary results. They differ from scalar functions in that they take grouped sets of relational data, make a pass over each group, and return one result for the group. Some examples of standard SQL aggregate functions are AVG, SUM, MAX, and MIN. intermediate numeric numeric ** numeric (exponentiation) numeric numeric * numeric (multiplication) numeric / numeric (division) numeric MOD numeric (modulo operator) numeric numeric + numeric (addition) numeric - numeric (subtraction) string concatenation operator logical value EQ value value NE value value GT value value LE value value LT value value GE value value IN set value NOT IN set value BETWEEN value AND value character value LIKE character value logical NOT logical logical logical AND logical lowest logical logical OR logical Precedence Result Type OperationChapter 2: Basic SQL Syntax and Lexicon Delimiters SQL Reference: Fundamentals 93 Related Topics For the names, parameters, return values, and other details of scalar and aggregate functions, see SQL Reference: Functions and Operators. Delimiters Introduction Delimiters are special characters having meanings that depend on context. The function of each delimiter appears in the following table. Delimiter Name Purpose ( ) LEFT PARENTHESIS RIGHT PARENTHESIS Group expressions and define the limits of various phrases. , COMMA Separates and distinguishes column names in the select list, or column names or parameters in an optional clause, or DateTime fields in a DateTime type. : COLON Prefixes reference parameters or client system variables. Also separates DateTime fields in a DateTime type. . FULLSTOP • Separates database names from table, trigger, UDF, UDT, and stored procedure names, such as personnel.employee. • Separates table names from a particular column name, such as employee.deptno). • In numeric constants, the period is the decimal point. • Separates DateTime fields in a DateTime type. • Separates a method name from a UDT expression in a method invocation. ; SEMICOLON • Separates statements in multi-statement requests. • Separates statements in a stored procedure body. • Separates SQL procedure statements in a triggered SQL statement in a trigger definition. • Terminates requests submitted via utilities such as BTEQ. • Terminates embedded SQL statements in C or PL/I applications. ’ APOSTROPHE • Defines the boundaries of character string constants. • To include an APOSTROPHE character or show possession in a title, double the APOSTROPHE characters. • Also separates DateTime fields in a DateTime type.Chapter 2: Basic SQL Syntax and Lexicon Separators 94 SQL Reference: Fundamentals Example In the following statement submitted through BTEQ, the FULLSTOP separates the database name (Examp and Personnel) from the table name (Profile and Employee), and, where reference is qualified to avoid ambiguity, it separates the table name (Profile, Employee) from the column name (DeptNo). UPDATE Examp.Profile SET FinGrad = 'A' WHERE Name = 'Phan A' ; SELECT EdLev, FinGrad,JobTitle, YrsExp FROM Examp.Profile, Personnel.Employee WHERE Profile.DeptNo = Employee.DeptNo ; The first SEMICOLON separates the UPDATE statement from the SELECT statement. The second SEMICOLON terminates the entire multistatement request. The semicolon is required in Teradata SQL to separate multiple statements in a request and to terminate a request submitted through BTEQ. Separators Lexical Separators A lexical separator is a character string that can be placed between words, literals, and delimiters without changing the meaning of a statement. Valid lexical separators are any of the following. • Comments For an explanation of comment lexical separators, see “Comments” on page 95. • Pad characters (several pad characters are treated as a single pad character except in a string literal) • RETURN characters (X’0D’) Statement Separators The SEMICOLON is a Teradata SQL statement separator. “ QUOTATION MARK Defines the boundaries of nonstandard names. / SOLIDUS Separates DateTime fields in a DateTime type. B b Uppercase B Lowercase b - HYPHENMINUS Delimiter Name PurposeChapter 2: Basic SQL Syntax and Lexicon Comments SQL Reference: Fundamentals 95 Each statement of a multistatement request must be separated from any subsequent statement with a semicolon. The following multistatement request illustrates the use of the semicolon as a statement separator. SHOW TABLE Payroll_Test ; INSERT INTO Payroll_Test (EmpNo, Name, DeptNo) VALUES ('10044', 'Jones M', '300') ; INSERT INTO ... For statements entered using BTEQ, a request terminates with an input line-ending semicolon unless that line has a comment, beginning with two dashes (- -). Everything to the right of the - - is a comment. In this case, the semicolon must be on the following line. The SEMICOLON as a statement separator in a multistatement request is a Teradata extension to the ANSI SQL-2003 standard. Comments Introduction You can embed comments within an SQL request anywhere a blank can occur. The SQL parser and the preprocessor recognize the following types of embedded comments: • Simple • Bracketed Simple Comments The simple form of a comment is delimited by two consecutive HYPHEN-MINUS (U+002D) characters (--) at the beginning of the comment and the newline character at the end of the comment. The newline character is implementation-specific, but is typed by pressing the Enter (non- 3270 terminals) or Return (3270 terminals) key. Simple SQL comments cannot span multiple lines. Example The following SELECT statement illustrates the use of a simple comment: SELECT EmpNo, Name FROM Payroll_Test ORDER BY Name -- Alphabetic order ; 1101E231 - - comment_text new_line_characterChapter 2: Basic SQL Syntax and Lexicon Terminators 96 SQL Reference: Fundamentals Bracketed Comments A bracketed comment is a text string of unrestricted length that is delimited by the beginning comment characters SOLIDUS (U+002F) and ASTERISK (U+002A) /* and the end comment characters ASTERISK and SOLIDUS */. Bracketed comments can begin anywhere on an input line and can span multiple lines. Example The following CREATE TABLE statement illustrates the use of a bracketed comment. CREATE TABLE Payroll_Test /* This is a test table set up to process actual payroll data on a test basis. The data generated from this table will be compared with the existing payroll system data for 2 months as a parallel test. */ (EmpNo INTEGER NOT NULL FORMAT ’ZZZZ9’, Name VARCHAR(12) NOT NULL, DeptNo INTEGER FORMAT ’ZZZZ9’, . . . Comments With Multibyte Character Set Strings You can include multibyte character set strings in both simple and bracketed comments. When using mixed mode in comments, you must have a properly formed mixed mode string, which means that a Shift-In (SI) must follow its associated Shift-Out (SO). If an SI does not follow the multibyte string, the results are unpredictable. When using bracketed comments that span multiple lines, the SI must be on the same line as its associated SO. If the SI and SO are not on the same line, the results are unpredictable. You must specify the bracketed comment delimiters (/* and */) as single byte characters. Terminators Definition The SEMICOLON is a Teradata SQL request terminator when it is the last non-blank character on an input line in BTEQ unless that line has a comment beginning with two dashes. In this case, the SEMICOLON request terminator should be on the following line, after the comment line. /* comment_text */ 1101E230Chapter 2: Basic SQL Syntax and Lexicon Terminators SQL Reference: Fundamentals 97 A request is considered complete when either the “End of Text” character or the request terminator character is detected. ANSI Compliance The SEMICOLON as a request terminator is Teradata extension to the ANSI SQL-2003 standard. Example For example, on the following input line: SELECT * FROM Employee ; the SEMICOLON terminates the single-statement request “SELECT * FROM Employee”. BTEQ uses SEMICOLONs to terminate multistatement requests. A request terminator is mandatory for request types that are: • In the body of a macro • Triggered action statements in a trigger definition • Entered using the BTEQ interface • Entered using other interfaces that require BTEQ Example 1: Macro Request The following statement illustrates the use of a request terminator in the body of a macro. CREATE MACRO Test_Pay (number (INTEGER), name (VARCHAR(12)), dept (INTEGER) AS ( INSERT INTO Payroll_Test (EmpNo, Name, DeptNo) VALUES (:number, :name, :dept) ; UPDATE DeptCount SET EmpCount = EmpCount + 1 ; SELECT * FROM DeptCount ; ) Example 2: BTEQ Request When entered through BTEQ, the entire CREATE MACRO statement must be terminated. CREATE MACRO Test_Pay (number (INTEGER), name (VARCHAR(12)), dept (INTEGER) AS (INSERT INTO Payroll_Test (EmpNo, Name, DeptNo) VALUES (:number, :name, :dept) ; UPDATE DeptCount SET EmpCount = EmpCount + 1 ; SELECT * FROM DeptCount ; ) ;Chapter 2: Basic SQL Syntax and Lexicon Null Statements 98 SQL Reference: Fundamentals Null Statements Introduction A null statement is a statement that has no content except for optional pad characters or SQL comments. Example 1 The semicolon in the following request is a null statement. /* This example shows a comment followed by a semicolon used as a null statement */ ; UPDATE Pay_Test SET ... Example 2 The first SEMICOLON in the following request is a null statement. The second SEMICOLON is taken as statement separator: /* This example shows a semicolon used as a null statement and as a statement separator */ ; UPDATE Payroll_Test SET Name = 'Wedgewood A' WHERE Name = 'Wedgewood A' ; SELECT ... -- This example shows the use of an ANSI component -- used as a null statement and statement separator ; Example 3 A SEMICOLON that precedes the first (or only) statement of a request is taken as a null statement. ;DROP TABLE temp_payroll;SQL Reference: Fundamentals 99 CHAPTER 3 SQL Data Definition, Control, and Manipulation This chapter describes the functional families of the SQL language. Topics include: • SQL Functional Families and Binding Styles • Data Definition Language • Data Control Language • Data Manipulation Language • Query and Workload Analysis Statements • Help and Database Object Definition Tools SQL Functional Families and Binding Styles Introduction The SQL language can be characterized in several different ways. This chapter is organized around functional groupings of the components of the language with minor emphasis on binding styles. Definition: Functional Family SQL provides facilities for defining database objects, for defining user access to those objects, and for manipulating the data stored within them. The following list describes the principal functional families of the SQL language. • SQL Data Definition Language (DDL) • SQL Data Control Language (DCL) • SQL Data Manipulation Language (DML) • Query and Workload Analysis Statements • Help and Database Object Definition Tools Some classifications of SQL group the data control language statements with the data definition language statements.Chapter 3: SQL Data Definition, Control, and Manipulation Embedded SQL 100 SQL Reference: Fundamentals Definition: Binding Style The ANSI SQL standards do not define the term binding style. The expression refers to a possible method by which an SQL statement can be invoked. Teradata Database supports the following SQL binding styles: • Direct, or interactive • Embedded SQL • Stored procedure • SQL Call Level Interface (as ODBC) • JDBC The direct binding style is usually not qualified in this manual set because it is the default style. Embedded SQL and stored procedure binding styles are always clearly specified, either explicitly or by context. Related Topics You can find more information on binding styles in the SQL Reference set and in other books. Embedded SQL You can execute SQL statements from within client application programs. The expression embedded SQL refers to SQL statements executed or declared from within a client application. An embedded Teradata SQL client program consists of the following: • Client programming language statements • One or more embedded SQL statements • Depending on the host language, one or more embedded SQL declare sections SQL declare sections are optional in COBOL and PL/I, but must be used in C. FOR more information on … SEE … embedded SQL • “Embedded SQL” on page 100 • Teradata Preprocessor2 for Embedded SQL Programmer Guide • SQL Reference: Stored Procedures and Embedded SQL stored procedures • “Stored Procedures” on page 48 • SQL Reference: Stored Procedures and Embedded SQL ODBC ODBC Driver for Teradata User Guide JDBC Teradata Driver for the JDBC Interface User GuideChapter 3: SQL Data Definition, Control, and Manipulation Data Definition Language SQL Reference: Fundamentals 101 A special prefix, EXEC SQL, distinguishes the SQL language statements embedded into the application program from the host programming language. Embedded SQL statements must follow the rules of the host programming language concerning statement continuation and termination, construction of variable names, and so forth. Aside from these rules, embedded SQL is host language-independent. Details of Teradata Database support for embedded SQL are described in SQL Reference: Stored Procedures and Embedded SQL. Data Definition Language Definition The SQL Data Definition Language (DDL) is a subset of the SQL language and consists of all SQL statements that support the definition of database objects. Purpose of Data Definition Language Statements Data definition language statements perform the following functions: • Create, drop, rename, and alter tables • Create, drop, rename, and replace stored procedures, user-defined functions, views, and macros • Create, drop, and alter user-defined types • Create, drop, and replace user-defined methods • Create and drop indexes • Create, drop, and modify users and databases • Create, drop, alter, rename, and replace triggers • Create, drop, and set roles • Create, drop, and modify profiles • Collect statistics on a column set or index • Establish a default database • Comment on database objects • Set a different collation sequence, account priority, DateForm, time zone, and database for the session • Begin and end logging Rules on Entering DDL Statements A DDL statement can be entered as: • A single statement request. • The solitary statement, or the last statement, in an explicit transaction (in Teradata mode, one or more requests enclosed by user-supplied BEGIN TRANSACTION and END Chapter 3: SQL Data Definition, Control, and Manipulation Data Definition Language 102 SQL Reference: Fundamentals TRANSACTION statement, or in ANSI mode, one or more requests ending with the COMMIT keyword). • The solitary statement in a macro. DDL statements cannot be entered as part of a multistatement request. Successful execution of a DDL statement automatically creates and updates entries in the Data Dictionary. SQL Data Definition Statements DDL statements include the following: Related Topics For detailed information about the function, syntax, and usage of Teradata SQL Data Definition statements, see SQL Reference: Data Definition Statements. • ALTER FUNCTION • ALTER METHOD • ALTER PROCEDURE • ALTER REPLICATION GROUP • ALTER TABLE • ALTER TRIGGER • ALTER TYPE • BEGIN LOGGING • COMMENT • CREATE AUTHORIZATION • CREATE CAST • CREATE DATABASE • CREATE FUNCTION • CREATE HASH INDEX • CREATE INDEX • CREATE JOIN INDEX • CREATE MACRO • CREATE METHOD • CREATE ORDERING • CREATE PROCEDURE • CREATE PROFILE • CREATE REPLICATION GROUP • CREATE ROLE • CREATE TABLE • CREATE TRANSFORM • CREATE TRIGGER • CREATE TYPE • CREATE USER • CREATE VIEW • DATABASE • DELETE DATABASE • DELETE USER • DROP AUTHORIZATION • DROP CAST • DROP DATABASE • DROP FUNCTION • DROP HASH INDEX • DROP INDEX • DROP JOIN INDEX • DROP MACRO • DROP ORDERING • DROP PROCEDURE • DROP PROFILE • DROP REPLICATION GROUP • DROP ROLE • DROP TABLE • DROP TRANSFORM • DROP TRIGGER • DROP TYPE • DROP USER • DROP VIEW • END LOGGING • MODIFY DATABASE • MODIFY PROFILE • MODIFY USER • RENAME FUNCTION • RENAME MACRO • RENAME PROCEDURE • RENAME TABLE • RENAME TRIGGER • RENAME VIEW • REPLACE CAST • REPLACE FUNCTION • REPLACE MACRO • REPLACE METHOD • REPLACE ORDERING • REPLACE PROCEDURE • REPLACE TRANSFORM • REPLACE TRIGGER • REPLACE VIEW • SET ROLE • SET SESSION • SET TIME ZONEChapter 3: SQL Data Definition, Control, and Manipulation Altering Table Structure and Definition SQL Reference: Fundamentals 103 Altering Table Structure and Definition Introduction You may need to change the structure or definition of an existing table or temporary table. In many cases, you can use ALTER TABLE and RENAME to make the changes. Some changes, however, may require you to use CREATE TABLE to recreate the table. How to Make Changes Use the RENAME TABLE statement to change the name of a table or temporary table. Use the ALTER TABLE statement to perform any of the following functions: • Add or drop columns on an existing table or temporary table • Add column default control, FORMAT, and TITLE attributes on an existing table or temporary table • Add or remove journaling options on an existing table or temporary table • Add or remove the FALLBACK option on an existing table or temporary table • Change the DATABLOCKSIZE or percent FREESPACE on an existing table or temporary table • Add or drop column and table level constraints on an existing table or temporary table • Change the LOG and ON COMMIT options for a global temporary table • Modify referential constraints • Change the properties of the primary index for a table (some cases require an empty table) • Change the partitioning properties of the primary index for a table, including modifications to the partitioning expression defined for use by a partitioned primary index (some cases require an empty table) • Regenerate table headers and optionally validate and correct the partitioning of PPI table rows • Define, modify, or delete the COMPRESS attribute for an existing column • Change column attributes (that do not affect stored data) on an existing table or temporary table Restrictions apply to many of the preceding modifications. For a complete list of rules and restrictions on using ALTER TABLE to change the structure or definition of an existing table, see SQL Reference: Data Definition Statements. To perform any of the following functions, use CREATE TABLE to recreate the table: • Redefine the primary index or its partitioning for a non-empty table when not allowed for ALTER TABLE • Change a data type attribute that affects existing data • Add a column that would exceed the maximum column count Interactively, the SHOW TABLE statement can call up the current table definition, which can then be modified and resubmitted to create a new table.Chapter 3: SQL Data Definition, Control, and Manipulation Dropping and Renaming Objects 104 SQL Reference: Fundamentals If the stored data is not affected by incompatible data type changes, an INSERT... SELECT statement can be used to transfer data from the existing table to the new table. Dropping and Renaming Objects Dropping Objects To drop an object, use the appropriate DDL statement. Renaming Objects Teradata SQL provides RENAME statements that you can use to rename some objects. To rename objects that do not have associated RENAME statements, you must first drop them and then recreate them with a new name, or, in the case of primary indexes, use ALTER TABLE. To drop this type of database object … Use this SQL statement … Hash Index DROP HASH INDEX Join Index DROP JOIN INDEX Macro DROP MACRO Profile DROP PROFILE Role DROP ROLE Secondary Index DROP INDEX Stored procedure DROP PROCEDURE Table DROP TABLE Global temporary table or volatile table Primary index Trigger DROP TRIGGER User-Defined Function DROP FUNCTION User-Defined Method ALTER TYPE User-Defined Type DROP TYPE View DROP VIEW To rename this type of database object … Use … Hash index DROP HASH INDEX and then CREATE HASH INDEX Join index DROP JOIN INDEX and then CREATE JOIN INDEXChapter 3: SQL Data Definition, Control, and Manipulation Data Control Language SQL Reference: Fundamentals 105 Related Topics For further information on these statements, including rules that apply to usage, see SQL Reference: Data Definition Statements. Data Control Language Definition The SQL Data Control Language (DCL) is a subset of the SQL language and consists of all SQL statements that support the definition of security authorization for accessing database objects. Purpose of Data Control Statements Data control statements perform the following functions: • Grant and revoke privileges • Give ownership of a database to another user Rules on Entering Data Control Statements A data control statement can be entered as: • A single statement request • The solitary statement, or as the last statement, in an “explicit transaction” (one or more requests enclosed by user-supplied BEGIN TRANSACTION and END TRANSACTION Macro RENAME MACRO Primary index ALTER TABLE Profile DROP PROFILE and then CREATE PROFILE Role DROP ROLE and then CREATE ROLE Secondary index DROP INDEX and then CREATE INDEX Stored procedure RENAME PROCEDURE Table RENAME TABLE Global temporary table or volatile table Trigger RENAME TRIGGER User-Defined Function RENAME FUNCTION User-Defined Method ALTER TYPE and then CREATE METHOD User-Defined Type DROP TYPE and then CREATE TYPE View RENAME VIEW To rename this type of database object … Use …Chapter 3: SQL Data Definition, Control, and Manipulation Data Manipulation Language 106 SQL Reference: Fundamentals statement in Teradata mode, or in ANSI mode, one or more requests ending with the COMMIT keyword). • The solitary statement in a macro A data control statement cannot be entered as part of a multistatement request. Successful execution of a data control statement automatically creates and updates entries in the Data Dictionary. Teradata SQL Data Control Statements Data control statements include the following: • GIVE • GRANT • GRANT LOGON • REVOKE • REVOKE LOGON Related Topics For detailed information about the function, syntax, and usage of Teradata SQL Data Control statements, see “SQL Data Control Language Statement Syntax” in SQL Reference: Data Definition Statements. Data Manipulation Language Definition The SQL Data Manipulation Language (DML) is a subset of the SQL language and consists of all SQL statements that support the manipulation or processing of database objects. Selecting Columns The SELECT statement returns information from the tables in a relational database. SELECT specifies the table columns from which to obtain the data, the corresponding database (if not defined by default), and the table (or tables) to be accessed within that database. For example, to request the data from the name, salary, and jobtitle columns of the Employee table, type: SELECT name, salary, jobtitle FROM employee ; The response might be something like the following results table. Chapter 3: SQL Data Definition, Control, and Manipulation Data Manipulation Language SQL Reference: Fundamentals 107 Note: The left-to-right order of the columns in a result table is determined by the order in which the column names are entered in the SELECT statement. Columns in a relational table are not ordered logically. As long as a statement is otherwise constructed properly, the spacing between statement elements is not important as long as at least one pad character separates each element that is not otherwise separated from the next. For example, the SELECT statement in the above example could just as well be formulated like this: SELECT name, salary,jobtitle FROM employee; Notice that there are multiple pad characters between most of the elements and that a comma only (with no pad characters) separates column name salary from column name jobtitle. To select all the data in the employee table, you could enter the following SELECT statement: SELECT * FROM employee ; The asterisk specifies that the data in all columns (except system-derived columns) of the table is to be returned. Selecting Rows The SELECT statement retrieves stored data from a table. All rows, specified rows, or specific columns of all or specified rows can be retrieved. The FROM, WHERE, ORDER BY, DISTINCT, WITH, GROUP BY, HAVING, and TOP clauses provide for a fine detail of selection criteria. To obtain data from specific rows of a table, use the WHERE clause of the SELECT statement. That portion of the clause following the keyword WHERE causes a search for rows that satisfy the condition specified. For example, to get the name, salary, and title of each employee in Department 100, use the WHERE clause: SELECT name, salary, jobtitle FROM employee WHERE deptno = 100 ; Name Salary JobTitle Newman P 28600.00 Test Tech Chin M 38000.00 Controller Aquilar J 45000.00 Manager Russell S 65000.00 President Clements D 38000.00 SalespersonChapter 3: SQL Data Definition, Control, and Manipulation Data Manipulation Language 108 SQL Reference: Fundamentals The response appears in the following table. To obtain data from a multirow result table in embedded SQL, declare a cursor for the SELECT statement and use it to fetch individual result rows for processing. To obtain data from the row with the oldest timestamp value in a queue table, use the SELECT AND CONSUME statement, which also deletes the row from the queue table. Zero-Table SELECT Zero-table SELECT statements return data but do not access tables. For example, the following SELECT statement specifies an expression after the SELECT keyword that does not require a column reference or FROM clause: SELECT 40000.00 / 52.; The response is one row: (40000.00/52.) ----------------- 769.23 Here is another example that specifies an attribute function after the SELECT keyword: SELECT TYPE(sales_table.region); Because the argument to the TYPE function is a column reference that specifies the table name, a FROM clause is not required and the query does not access the table. The response is one row that might be something like the following: Type(region) --------------------------------------- INTEGER Adding Rows Use the INSERT statement to add rows to a table. One statement is required for each new row, except in the case of an INSERT...SELECT statement. For more details on this, see SQL Reference: Data Manipulation Statements. Defaults and constraints defined by the CREATE TABLE statement affect an insert operation in the following ways. Name Salary JobTitle Chin M 38000.00 Controller Greene W 32500.00 Payroll Clerk Moffit H 35000.00 Recruiter Peterson J 25000.00 Payroll ClerkChapter 3: SQL Data Definition, Control, and Manipulation Data Manipulation Language SQL Reference: Fundamentals 109 Updating Rows To modify data in one or more rows of a table, use the UPDATE statement. In the UPDATE statement, you specify the column name of the data to be modified along with the new value. You can also use a WHERE clause to qualify the rows to change. Attributes specified in the CREATE TABLE statement affect an update operation in the following ways: • When an update supplies a value that violates some defined constraint on a column or columns, the update operation is rejected and an error message is returned. • When an update supplies the value NULL and a NULL is allowed, any existing data is removed from the column. • If the result of an UPDATE will violate uniqueness constraints or create a duplicate row in a table which does not allow duplicate rows, an error message is returned. To update rows in a multirow result table in embedded SQL, declare a cursor for the SELECT statement and use it to fetch individual result rows for processing, then use a WHERE CURRENT OF clause in a positioned UPDATE statement to update the selected rows. The Teradata Database supports a special form of UPDATE, called the upsert form, which is a single SQL statement that includes both UPDATE and INSERT functionality. The specified update operation performs first, and if it fails to find a row to update, then the specified insert operation performs automatically. Alternatively, use the MERGE statement. Deleting Rows The DELETE statement allows you to remove an entire row or rows from a table. A WHERE clause qualifies the rows that are to be deleted. WHEN an INSERT statement … THEN the system … attempts to add a duplicate row • for any unique index • to a table defined as SET (not to allow duplicate rows) returns an error, with one exception. The system silently ignores duplicate rows that an INSERT … SELECT would create when the: • table is defined as SET • mode is Teradata omits a value for a column for which a default value is defined stores the default value for that column. omits a value for a column for which both of the following statements are true: • NOT NULL is specified • no default is specified rejects the operation and returns an error message. supplies a value that does not satisfy the constraints specified for a column or violates some defined constraint on a column or columnsChapter 3: SQL Data Definition, Control, and Manipulation Subqueries 110 SQL Reference: Fundamentals To delete rows in a multirow result table in embedded SQL, use the following process: 1 Declare a cursor for the SELECT statement. 2 Fetch individual result rows for processing using the cursor you declared. 3 Use a WHERE CURRENT OF clause in a positioned DELETE statement to delete the selected rows. Merging Rows The MERGE statement merges a source row into a target table based on whether any target rows satisfy a specified matching condition with the source row. The MERGE statement is a single SQL statement that includes both UPDATE and INSERT functionality. Subqueries Introduction Subqueries are nested SELECT statements. They can be used to ask a series of questions to arrive at a single answer. Three Level Subqueries: Example The following subqueries, nested to three levels, are used to answer the question “Who manages the manager of Marston?” SELECT Name FROM Employee WHERE EmpNo IN (SELECT MgrNo FROM Department WHERE DeptNo IN (SELECT DeptNo FROM Employee WHERE Name = 'Marston A') ) ; The subqueries that pose the questions leading to the final answer are inverted: • The third subquery asks the Employee table for the number of Marston’s department. • The second subquery asks the Department table for the employee number (MgrNo) of the manager associated with this department number. • The first subquery asks the Employee table for the name of the employee associated with this employee number (MgrNo). IF the source and target rows … THEN the merge operation is an … satisfy the matching condition update based on the specified WHEN MATCHED THEN UPDATE clause. do not satisfy the matching condition insert based on the specified WHEN NOT MATCHED THEN INSERT clause.Chapter 3: SQL Data Definition, Control, and Manipulation Recursive Queries SQL Reference: Fundamentals 111 The result table looks like the following: Name -------- Watson L This result can be obtained using only two levels of subquery, as the following example shows. SELECT Name FROM Employee WHERE EmpNo IN (SELECT MgrNo FROM Department, Employee WHERE Employee.Name = 'Marston A' AND Department.DeptNo = Employee.DeptNo) ; In this example, the second subquery defines a join of Employee and Department tables. This result could also be obtained using a one-level query that uses correlation names, as the following example shows. SELECT M.Name FROM Employee M, Department D, Employee E WHERE M.EmpNo = D.MgrNo AND E.Name = 'Marston A' AND D.DeptNo = E.DeptNo; In some cases, as in the preceding example, the choice is a style preference. In other cases, correct execution of the query may require a subquery. For More Information For more information, see SQL Reference: Data Manipulation Statements. Recursive Queries Introduction A recursive query is a way to query hierarchies of data, such as an organizational structure, bill-of-materials, and document hierarchy. Recursion is typically characterized by three steps: 1 Initialization 2 Recursion, or repeated iteration of the logic through the hierarchy 3 Termination Similarly, a recursive query has three execution phases: 1 Create an initial result set. 2 Recursion based on the existing result set. 3 Final query to return the final result set.Chapter 3: SQL Data Definition, Control, and Manipulation Recursive Queries 112 SQL Reference: Fundamentals Two Ways to Specify a Recursive Query You can specify a recursive query by: • Preceding a query with the WITH RECURSIVE clause • Creating a permanent view using the RECURSIVE clause in a CREATE VIEW statement Using the WITH RECURSIVE Clause Consider the following employee table: CREATE TABLE employee (employee_number INTEGER ,manager_employee_number INTEGER ,last_name CHAR(20) ,first_name VARCHAR(30)); The table represents an organizational structure containing a hierarchy of employee-manager data. The following figure depicts what the employee table looks like hierarchically. The following recursive query retrieves the employee numbers of all employees who directly or indirectly report to the manager with employee_number 801: WITH RECURSIVE temp_table (employee_number) AS ( SELECT root.employee_number FROM employee root WHERE root.manager_employee_number = 801 UNION ALL SELECT indirect.employee_number FROM temp_table direct, employee indirect 1101A285 employee # = 801 manager employee # = NULL employee # = 1003 manager employee # = 801 employee # = 1010 manager employee # = 1003 employee # = 1012 manager employee # = 1004 employee # = 1002 manager employee # = 1004 employee # = 1015 manager employee # = 1004 employee # = 1001 manager employee # = 1003 employee # = 1004 manager employee # = 1003 employee # = 1008 manager employee # = 1019 employee # = 1006 manager employee # = 1019 employee # = 1014 manager employee # = 1019 employee # = 1011 manager employee # = 1019 employee # = 1019 manager employee # = 801 employee # = 1016 manager employee # = 801Chapter 3: SQL Data Definition, Control, and Manipulation Recursive Queries SQL Reference: Fundamentals 113 WHERE direct.employee_number = indirect.manager_employee_number ) SELECT * FROM temp_table ORDER BY employee_number; In the example, temp_table is a temporary named result set that can be referred to in the FROM clause of the recursive statement. The initial result set is established in temp_table by the non-recursive, or seed, statement and contains the employees that report directly to the manager with an employee_number of 801: SELECT root.employee_number FROM employee root WHERE root.manager_employee_number = 801 The recursion takes place by joining each employee in temp_table with employees who report to the employees in temp_table. The UNION ALL adds the results to temp_table. SELECT indirect.employee_number FROM temp_table direct, employee indirect WHERE direct.employee_number = indirect.manager_employee_number Recursion stops when no new rows are added to temp_table. The final query is not part of the recursive WITH clause and extracts the employee information out of temp_table: SELECT * FROM temp_table ORDER BY employee_number; Here are the results of the recursive query: employee_number --------------- 1001 1002 1003 1004 1006 1008 1010 1011 1012 1014 1015 1016 1019 Using the RECURSIVE Clause in a CREATE VIEW Statement Creating a permanent view using the RECURSIVE clause is similar to preceding a query with the WITH RECURSIVE clause. Consider the employee table that was presented in “Using the WITH RECURSIVE Clause” on page 112. The following statement creates a view named hierarchy_801 using a recursive query that retrieves the employee numbers of all employees who directly or indirectly report to the manager with employee_number 801: CREATE RECURSIVE VIEW hierarchy_801 (employee_number) AS ( SELECT root.employee_number FROM employee rootChapter 3: SQL Data Definition, Control, and Manipulation Recursive Queries 114 SQL Reference: Fundamentals WHERE root.manager_employee_number = 801 UNION ALL SELECT indirect.employee_number FROM hierarchy_801 direct, employee indirect WHERE direct.employee_number = indirect.manager_employee_number ); The seed statement and recursive statement in the view definition are the same as the seed statement and recursive statement in the previous recursive query that uses the WITH RECURSIVE clause, except that the hierarchy_801 view name is different from the temp_table temporary result name. To extract the employee information, use the following SELECT statement on the hierarchy_801 view: SELECT * FROM hierarchy_801 ORDER BY employee_number; Here are the results: employee_number --------------- 1001 1002 1003 1004 1006 1008 1010 1011 1012 1014 1015 1016 1019 Depth Control to Avoid Infinite Recursion If the hierarchy is cyclic, or if the recursive statement specifies a bad join condition, a recursive query can produce a runaway query that never completes with a finite result. The best practice is to control the depth of the recursion as follows: • Specify a depth control column in the column list of the WITH RECURSIVE clause or recursive view • Initialize the column value to 0 in the seed statements • Increment the column value by 1 in the recursive statements • Specify a limit for the value of the depth control column in the join condition of the recursive statements Here is an example that modifies the previous recursive query that uses the WITH RECURSIVE clause of the employee table to limit the depth of the recursion to five cycles: WITH RECURSIVE temp_table (employee_number, depth) AS ( SELECT root.employee_number, 0 AS depth FROM employee root WHERE root.manager_employee_number = 801 Chapter 3: SQL Data Definition, Control, and Manipulation Query and Workload Analysis Statements SQL Reference: Fundamentals 115 UNION ALL SELECT indirect.employee_number, direct.depth+1 AS newdepth FROM temp_table direct, employee indirect WHERE direct.employee_number = indirect.manager_employee_number AND newdepth <= 5 ) SELECT * FROM temp_table ORDER BY employee_number; Related Topics Query and Workload Analysis Statements Data Collection and Analysis Collected data can be used in several ways, for example: • By the Optimizer, to produce the best query plans possible. • To populate user-defined Query Capture Database (QCD) tables with data used by various utilities to analyze query workloads as part of the ongoing process to reengineer the database design process. For example, the Teradata Index Wizard determines optimal secondary index sets to support the query workloads you ask it to analyze. Index Analysis and Target Level Emulation Teradata also provides diagnostic statements that support the Teradata Index Wizard and the sample-based components of the target level emulation facility used to emulate a production environment on a test system: • DIAGNOSTIC DUMP SAMPLES • DIAGNOSTIC HELP SAMPLES FOR details on … SEE … recursive queries “WITH RECURSIVE” in SQL Reference: Data Manipulation Statements. recursive views “CREATE VIEW” in SQL Reference: Data Definition Statements. Teradata provides the following SQL statements for collecting and analyzing query and data demographics and statistics: • BEGIN QUERY LOGGING • COLLECT DEMOGRAPHICS • COLLECT STATISTICS • DROP STATISTICS • DUMP EXPLAIN • END QUERY LOGGING • INITIATE INDEX ANALYSIS • INSERT EXPLAIN • RESTART INDEX ANALYSISChapter 3: SQL Data Definition, Control, and Manipulation Help and Database Object Definition Tools 116 SQL Reference: Fundamentals • DIAGNOSTIC SET SAMPLES • DIAGNOSTIC “Validate Index” After configuring the test environment and enabling it with the appropriate production system statistical and demographic data, you can perform various workload analyses to determine optimal sets of secondary indexes to support those workloads in the production environment. Related Topics For more information on query and workload analysis statements, see SQL Reference: Data Definition Statements. Help and Database Object Definition Tools Introduction Teradata SQL provides several powerful tools to get help about database object definitions and summaries of database object definition statement text. HELP Statements The various HELP statements return reports about the current column definitions for named database objects. The reports returned by these statements can be useful to database designers who need to fine tune index definitions, column definitions (for example, changing data typing to eliminate the necessity of ad hoc conversions), and so on. IF you want to get … THEN use … the attributes of a column, including whether it is a single-column primary or secondary index and, if so, whether it is unique HELP COLUMN the attributes for a specific named constraint on a table HELP CONSTRAINT the attributes, sorted by object name, for all tables, views, join and hash indexes, stored procedures, user-defined functions, and macros in a specified database HELP DATABASE and HELP USER the specific function name, list of parameters, data types of the parameters, and any comments associated with the parameters of a userdefined function HELP FUNCTION the data types of the columns defined by a particular hash index HELP HASH INDEX the attributes for the indexes defined for a table or join index HELP INDEX the attributes of the columns defined by a particular join index HELP JOIN INDEX the attributes for the specified macro HELP MACRO the specific name, list of parameters, data types of the parameters, and any comments associated with the parameters of a user-defined method HELP METHODChapter 3: SQL Data Definition, Control, and Manipulation Help and Database Object Definition Tools SQL Reference: Fundamentals 117 SHOW Statements A SHOW statement returns a CREATE statement indicating the last data definition statement performed against the named database object. These statements are particularly useful for application developers who need to develop exact replicas of existing objects for purposes of testing new software. the attributes for the specified join index or table HELP TABLE the attribute and format parameters for each parameter of the procedure or just the creation time attributes for the specified procedure HELP PROCEDURE the attributes of the specified replication group and its member tables HELP REPLICATION GROUP the attributes for the specified trigger HELP TRIGGER information on the type, attributes, methods, cast, ordering, and transform of the specified user-defined type HELP TYPE the attributes for a specified view HELP VIEW the attributes for the requested volatile table HELP VOLATILE TABLE IF you want to get … THEN use … IF you want to get the data definition statement most recently used to create, replace, or modify a specified … THEN use … hash index SHOW HASH INDEX join index SHOW JOIN INDEX macro SHOW MACRO stored procedure or external stored procedure SHOW PROCEDURE table SHOW TABLE trigger SHOW TRIGGER user-defined function SHOW FUNCTION user-defined method SHOW METHOD user-defined type SHOW TYPE view SHOW VIEWChapter 3: SQL Data Definition, Control, and Manipulation Help and Database Object Definition Tools 118 SQL Reference: Fundamentals Example Consider the following definition for a table named department: CREATE TABLE department, FALLBACK (department_number SMALLINT ,department_name CHAR(30) NOT NULL ,budget_amount DECIMAL(10,2) ,manager_employee_number INTEGER ) UNIQUE PRIMARY INDEX (department_number) ,UNIQUE INDEX (department_name); To get the attributes for the table, use the HELP TABLE statement: HELP TABLE department; The HELP TABLE statement returns: Column Name Type Comment ------------------------------ ---- ------------------------- department_number I2 ? department_name CF ? budget_amount D ? manager_employee_number I ? To get the CREATE TABLE statement that defines the department table, use the SHOW TABLE statement: SHOW TABLE department; The SHOW TABLE statement returns: CREATE SET TABLE TERADATA_EDUCATION.department, FALLBACK, NO BEFORE JOURNAL, NO AFTER JOURNAL, CHECKSUM = DEFAULT (department_number SMALLINT, department_name CHAR(30) CHARACTER SET LATIN NOT CASESPECIFIC NOT NULL, budget_amount DECIMAL(10,2), manager_employee_number INTEGER) UNIQUE PRIMARY INDEX ( department_number ) UNIQUE INDEX ( department_name ); Related Topics For more information, see SQL Reference: Data Definition Statements.SQL Reference: Fundamentals 119 CHAPTER 4 SQL Data Handling This chapter describes the fundamentals of Teradata Database data handling. Topics include: • Requests • Transactions • Event processing • Session Parameters • Session Management • Return Codes Invoking SQL Statements Introduction One of the primary issues that motivated the development of relational database management systems was the perceived need to create database management systems that could be queried not just by predetermined, hard-coded requests but also interactively by well-formulated ad hoc queries. SQL addresses this issue by offering four ways to invoke an executable statement: • Interactively from a terminal • Embedded within an application program • Dynamically performed from within an embedded application • Embedded within a stored procedure Executable SQL Statements An executable SQL statement is one that performs an action. The action can be on data or on a transaction or some other entity at a higher level than raw data. Some examples of executable SQL statements are the following: • SELECT • CREATE TABLE • COMMIT • CONNECT • PREPAREChapter 4: SQL Data Handling Requests 120 SQL Reference: Fundamentals Most, but not all, executable SQL statements can be performed interactively from a terminal using an SQL query manager like BTEQ or Teradata SQL Assistant (formerly called Queryman). Types of executable SQL commands that cannot be performed interactively are the following: • Cursor control and declaration statements • Dynamic SQL control statements • Stored procedure control statements and condition handlers • Connection control statements • Special forms of SQL statements such as SELECT INTO These statements can only be used within an embedded SQL or stored procedure application. Nonexecutable SQL Statements A nonexecutable SQL statement is one that declares an SQL statement, object, or host or local variable to the preprocessor or stored procedure compiler. Nonexecutable SQL statements are not processed during program execution. Some examples of nonexecutable SQL statements for embedded SQL applications include: • DECLARE CURSOR • BEGIN DECLARE SECTION • END DECLARE SECTION • EXEC SQL Examples of nonexecutable SQL statements for stored procedures include: • DECLARE CURSOR • DECLARE Requests Introduction A request to the Teradata Database consists of one or more SQL statements and can span any number of input lines. Teradata Database can receive and perform SQL statements that are: • Embedded in a client application program that is written in a procedural language. • Embedded in a stored procedure. • Entered interactively through BTEQ or Teradata SQL Assistant interfaces. • Submitted in a BTEQ script as a batch job. • Submitted through other supported methods (such as CLIv2, ODBC, and JDBC).Chapter 4: SQL Data Handling Requests SQL Reference: Fundamentals 121 Single Statement Requests A single statement request consists of a statement keyword followed by one or more expressions, other keywords, clauses, and phrases. A single statement request is treated as a solitary unit of work. Single Statement Syntax Multistatement Requests A multistatement request consists of two or more statements separated by SEMICOLON characters. Multistatement requests are non-ANSI standard. For more information, see “Multistatement Requests” on page 124. Multistatement Syntax Iterated Requests An iterated request is a single DML statement with multiple data records. Iterated requests do not directly impact the syntax of SQL statements. They provide a more performant way of processing DML statements that specify the USING row descriptor to import or export data from the Teradata Database. For more information, see “Iterated Requests” on page 127. ANSI Session Mode If an error is found in a request, then that request is aborted, but not the entire transaction. Note: Some failures will abort the entire transaction. Teradata Session Mode A multistatement request is treated as an implicit transaction. That is, if an error is found in any statement in the request, then the entire transaction is aborted. HH01A003 ; statement HH01A004 ; statement ;Chapter 4: SQL Data Handling Transactions 122 SQL Reference: Fundamentals Abort processing proceeds as follows: 1 Backs out any changes made to the database as a result of any preceding statements. 2 Deletes any associated spooled output. 3 Releases any associated locks. 4 Bypasses any remaining statements in the transaction. Complete Requests A request is considered complete when either an End of Text character or the request terminator is encountered. The request terminator is a SEMICOLON character. It is the last nonpad character on an input line. A request terminator is optional except when the request is embedded in an SQL macro or trigger or when it is entered through BTEQ. In a stored procedure, each SQL statement is treated as a request. Stored procedures do not support multistatement requests. Transactions Introduction A transaction is a logical unit of work where the statements nested within the transaction either execute successfully as a group or do not execute. Transaction Processing Mode You can perform transaction processing in either of the following session modes: • ANSI • Teradata In ANSI session mode, transaction processing adheres to the rules defined by the ANSI SQL specification. In Teradata session mode, transaction processing follows the rules defined by Teradata Database over years of evolution. To set the transaction processing mode, use the: • SessionMode field of the DBS Control Record • BTEQ command .SET SESSION TRANSACTION • Preprocessor2 TRANSACT() option • ODBC SessionMode option in the .odbc.ini file • JDBC TeraDataSource.setTransactMode() methodChapter 4: SQL Data Handling Transaction Processing in ANSI Session Mode SQL Reference: Fundamentals 123 Related Topics The next few pages highlight some of the differences between transaction processing in ANSI session mode and transaction processing in Teradata session mode. For detailed information on statement and transaction processing, see SQL Reference: Statement and Transaction Processing. Transaction Processing in ANSI Session Mode Introduction Transactions are always implicit in ANSI session mode. A transaction initiates when one of the following happens: • The first SQL statement in a session executes • The first statement following the close of a transaction executes The COMMIT or ROLLBACK/ABORT statements close a transaction. If a transaction includes a DDL statement, it must be the last statement in the transaction. Note that DATABASE and SET SESSION are DDL statements. See “Rollback Processing” in SQL Reference: Statement and Transaction Processing. If a session terminates with an open transaction, then any effects of that transaction are rolled back. Two-Phase Commit (2PC) Sessions in ANSI session mode do not support 2PC. If an attempt is made to use the 2PC protocol in ANSI session mode, the Logon process aborts and an error returns to the requestor. Transaction Processing in Teradata Session Mode Introduction A Teradata SQL transaction can be a single Teradata SQL statement, or a sequence of Teradata SQL statements, treated as a single unit of work. Each request is processed as one of the following transaction types: • Implicit • Explicit • Two-phase commit (2PC)Chapter 4: SQL Data Handling Multistatement Requests 124 SQL Reference: Fundamentals Implicit Transactions An implicit transaction is a request that does not include the BEGIN TRANSACTION and END TRANSACTION statements. The implicit transaction starts and completes all within the SQL request: it is self-contained. An implicit transaction can be one of the following: • A single DML statement that affects one or more rows of one or more tables • A macro or trigger containing one or more statements • A request containing multiple statements separated by SEMICOLON characters. Each SEMICOLON character can appear anywhere in the input line. The Parser interprets a SEMICOLON character at the end of an input line as a transaction terminator. DDL statements are not valid in an implicit multistatement transaction. Explicit Transactions In Teradata session mode, an explicit transaction contains one or more statements enclosed by BEGIN TRANSACTION and END TRANSACTION statements. The first BEGIN TRANSACTION initiates a transaction and the last END TRANSACTION terminates the transaction. When multiple statements are included in an explicit transaction, you can only specify a DDL statement if it is the last statement in the series. Two-Phase Commit (2PC) Rules Two-phase commit (2PC) protocol is supported in Teradata session mode: • A 2PC transaction contains one or more DML statements that affect multiple databases and are coordinated externally using the 2PC protocol. • A DDL statement is not valid in a two-phase commit transaction. Multistatement Requests Definition An atomic request containing more than one SQL statement, each terminated by a SEMICOLON character. Syntax HH01A004 ; statement ;Chapter 4: SQL Data Handling Multistatement Requests SQL Reference: Fundamentals 125 ANSI Compliance Multistatement requests are non-ANSI SQL-2003 standard. Rules The Teradata Database imposes restrictions on the use of multistatement requests: • Only one USING row descriptor is permitted per request, so only one USING row descriptor can be used per multistatement request. This rule applies to interactive SQL only. Embedded SQL and stored procedures do not permit the USING row descriptor. • A multistatement request cannot include a DDL statement. • The keywords BEGIN REQUEST and END REQUEST must delimit a multistatement request in a stored procedure. Power of Multistatement Requests The multistatement request is application-independent. It improves performance for a variety of applications that can package more than one SQL statement at a time. BTEQ, CLI, and the SQL preprocessor all support multistatement requests. Multistatement requests improve system performance by reducing processing overhead. By performing a series of statements as one request, performance for the client, the Parser, and the Database Manager are all enhanced. Because of this reduced overhead, using multistatement requests also decreases response time. A multistatement request that contains 10 SQL statements could be as much as 10 times more efficient than the 10 statements entered separately (depending on the types of statements submitted). Multistatement Requests Treated as Transaction In a multistatement request, treated as a single unit of work, either all statements in the request complete successfully, or the entire request is aborted. In ANSI session mode, the request is rolled back if aborted. In Teradata session mode, any updates to the database up to that point for the transaction are rolled back. Parallel Step Processing Teradata Database can perform some requests in parallel (see “Parallel Steps” on page 126). This capability applies both to implicit transactions, such as macros and multistatement requests, and to Teradata-style transactions explicitly defined by BEGIN/END TRANSACTION statements. Statements in a multistatement request are broken down by the Parser into one or more steps that direct the execution performed by the AMPs. It is these steps, not the actual statements, that are executed in parallel.Chapter 4: SQL Data Handling Multistatement Requests 126 SQL Reference: Fundamentals A handshaking protocol between the PE and the AMP allows the AMP to determine when the PE can dispatch the next parallel step. Up to twenty parallel steps can be processed per request if channels are not required, such as a request with an equality constraint based on a primary index value. Up to ten channels can be used for parallel processing when a request is not constrained to a primary index value. For example, if an INSERT step and a DELETE step are allowed to run in parallel, the AMP informs the PE that the DELETE step has progressed to the point where the INSERT step will not impact it adversely. This handshaking protocol also reduces the chance of a deadlock. “Parallel Steps” on page 126 illustrates the following process: 1 The statements in a multistatement request are broken down into a series of steps. 2 The Optimizer determines which steps in the series can be executed in parallel. 3 The steps are processed. Each step undergoes some preliminary processing before it is executed, such as placing locks on the objects involved. These preliminary processes are not performed in parallel with the steps. Parallel Steps Time Step 1 2 3 4 5 6 7 8 9 (1) Time (2) Time Step 1 2 5 6 9 7 8 (3) FF02A001 Step 1 2 3 4 5 6 7 8 9 3 4Chapter 4: SQL Data Handling Iterated Requests SQL Reference: Fundamentals 127 Iterated Requests Definition A single DML statement with multiple data records. Usage An iterated request is an atomic request consisting of a single SQL DML statement with multiple sets (records) of data. Iterated requests do not directly impact the syntax of SQL statements. They provide an efficient way to execute the same single-statement DML operation on multiple data records, like the way that ODBC applications execute parameterized statements for arrays of parameter values, for example. Several Teradata Database client tools and interfaces provide facilities to pack multiple data records in a single buffer with a single DML statement. For example, suppose you use BTEQ to import rows of data into table ptable using the following INSERT statement and USING row descriptor: USING (pid INTEGER, pname CHAR(12)) INSERT INTO ptable VALUES(:pid, :pname); To repeat the request as many times as necessary to read up to 200 data records and pack a maximum of 100 data records with each request, precede the INSERT statement with the following BTEQ command: .REPEAT RECS 200 PACK 100 Note: The PACK option is ignored if the database being used does not support iterated requests or if the request that follows the REPEAT command is not a DML statement supported by iterated requests. For details, see “Rules” on page 128. The following tools and interfaces provide facilities that you can use to execute iterated requests. Tool/Interface Facility CLIv2 for network-attached systems using_data_count field in the DBCAREA data area CLIv2 for channel-attached systems Using-data-count field in the DBCAREA data area ODBC Parameter arrays JDBC type 4 driver Batch operations OLE DB Provider for Teradata Parameter sets BTEQ • .REPEAT command • .SET PACK commandChapter 4: SQL Data Handling Iterated Requests 128 SQL Reference: Fundamentals Rules The following rules apply to iterated requests: • The iterated request must consist of a single DML statement from the following list: • ABORT • DELETE (excluding the positioned form of DELETE) • EXECUTE macro_name The fully-expanded macro must be equivalent to a single DML statement that is qualified to be in an iterated request. • INSERT • MERGE • ROLLBACK • SELECT • UPDATE (including atomic UPSERT, but excluding the positioned form of UPDATE) • The DML statement must reference user-supplied input data, either as named fields in a USING row descriptor or as '?' parameter markers in a parameterized request. • All the data records in a given request must use the same record layout. This restriction applies by necessity to requests where the record layout is given by a single USING row descriptor in the request text itself; but note that the restriction also applies to parameterized requests, where the request text has no USING descriptor and does not fully specify the input record. • The server processes the iterated request as if it were a single multi-statement request, with each iteration and its response associated with a corresponding statement number. Related Topics FOR more information on … SEE … iterated request processing SQL Reference: Statement and Transaction Processing which DML statements can be specified in an iterated request SQL Reference: Data Manipulation Statements CLIv2 • Teradata Call-Level Interface Version 2 Reference for Channel-Attached Systems • Teradata Call-Level Interface Version 2 Reference for Network-Attached Systems ODBC parameter arrays ODBC Driver for Teradata User Guide JDBC driver batch operations Teradata Driver for the JDBC Interface User Guide OLE DB Provider for Teradata parameter sets OLE DB Provider for Teradata Installation and User Guide BTEQ PACK command Basic Teradata Query ReferenceChapter 4: SQL Data Handling Dynamic and Static SQL SQL Reference: Fundamentals 129 Dynamic and Static SQL Definitions ANSI Compliance Dynamic SQL is ANSI SQL-2003-compliant. The ANSI SQL standard does not define the expression static SQL, but relational database management commonly uses it to contrast with the ANSI-defined expression dynamic SQL. Ad Hoc and Hard-Coded Invocation of SQL Statements Perhaps the best way to think of dynamic SQL is to contrast it with ad hoc SQL statements created and executed from a terminal and with preprogrammed SQL statements created by an application programmer and executed by an application program. In the case of the ad hoc query, everything legal is available to the requester: choice of SQL statements and clauses, variables and their names, databases, tables, and columns to manipulate, and literals. In the case of the application programmer, the choices are made in advance and hard-coded into the source code of the application. Once the program is compiled, nothing can be changed short of editing and recompiling the application. Dynamic Invocation of SQL Statements Dynamic SQL offers a compromise between these two extremes. By choosing to code dynamic SQL statements in the application, the programmer has the flexibility to allow an end user to select not only the variables to be manipulated at run time, but also the SQL statement to be executed. As you might expect, the flexibility that dynamic SQL offers a user is offset by more work and increased attention to detail on the part of the application programmer, who needs to set up additional dynamic SQL statements and manipulate information in the SQLDA to ensure a correct result. This is done by first preparing, or compiling, an SQL text string containing placeholder tokens at run time and then executing the prepared statement, allowing the application to prompt the user for values to be substituted for the placeholders. Term Definition Dynamic SQL Dynamic SQL is a method of invoking an SQL statement by compiling and performing it at runtime from within an embedded SQL application program or a stored procedure. The specification of data to be manipulated by the statement is also determined at runtime. Static SQL Static SQL is, by default, any method of invoking an SQL statement that is not dynamic.Chapter 4: SQL Data Handling Dynamic SQL in Stored Procedures 130 SQL Reference: Fundamentals SQL Statements to Set Up and Invoke Dynamic SQL The embedded SQL statements for preparing and executing an SQL statement dynamically are: • PREPARE • EXECUTE • EXECUTE IMMEDIATE. EXECUTE IMMEDIATE is a special form that combines PREPARE and EXECUTE into one statement. EXECUTE IMMEDIATE can only be used in the case where there are no input host variables. This description applies directly to all executable SQL statements except SELECT, which requires additional handling. Note that SELECT INTO cannot be invoked dynamically. For details, see SQL Reference: Stored Procedures and Embedded SQL. Related Topics Dynamic SQL in Stored Procedures Overview The way stored procedures support dynamic SQL statements is different from the way embedded SQL does. Use the following statement to set up and invoke dynamic SQL in a stored procedure: CALL DBC.SysExecSQL(string_expression) where string_expression is any valid string expression that builds an SQL statement. The string expression is composed of string literals, status variables, local variables, input (IN and INOUT) parameters, and for-loop aliases. Dynamic SQL statements are not validated at compile time. The resulting SQL statement cannot have status variables, local variables, parameters, for-loop aliases, or a USING or EXPLAIN modifier. For more information on … See … examples of dynamic SQL code in C, COBOL, and PL/I Teradata Preprocessor2 for Embedded SQL Programmer Guide.Chapter 4: SQL Data Handling Using SELECT With Dynamic SQL SQL Reference: Fundamentals 131 Example The following example uses dynamic SQL within stored procedure source text: CREATE PROCEDURE new_sales_table( my_table VARCHAR(30), my_database VARCHAR(30)) BEGIN DECLARE sales_columns VARCHAR(128) DEFAULT '(item INTEGER, price DECIMAL(8,2), sold INTEGER)'; CALL DBC.SysExecSQL('CREATE TABLE ' || my_database || '.' || my_table || sales_columns); END; Any number of calls to SysExecSQL can be made in a stored procedure and the request text in the string expression can specify a multistatement request. Because the request text of dynamic SQL statements can vary from execution to execution, dynamic SQL provides more usability and conciseness to the stored procedure definition. Restrictions Dynamic SQL statements can be specified in a stored procedure only when the creator is the same as the immediate "owner" of the stored procedure. The following SQL statements cannot be specified as dynamic SQL in stored procedures: Related Topics For rules and usage examples of dynamic SQL statements in stored procedures, see SQL Reference: Stored Procedures and Embedded SQL. Using SELECT With Dynamic SQL Unlike other executable SQL statements, SELECT returns information beyond statement responses and return codes to the requester. DESCRIBE Statement Because the requesting application needs to know how much (if any) data will be returned by a dynamically prepared SELECT, you must use an additional SQL statement, DESCRIBE, to make the application aware of the demographics of the data to be returned by the SELECT statement (see “DESCRIBE” in SQL Reference: Stored Procedures and Embedded SQL). • CALL • DATABASE • HELP • SELECT • SET SESSION ACCOUNT • SET SESSION DATEFORM • SHOW • CREATE PROCEDURE • EXPLAIN • REPLACE PROCEDURE • SELECT - INTO • SET SESSION COLLATION • SET TIME ZONEChapter 4: SQL Data Handling Using SELECT With Dynamic SQL 132 SQL Reference: Fundamentals DESCRIBE writes this information to the SQLDA declared for the SELECT statement as follows. General Procedure An application must use the following general procedure to set up, execute, and retrieve the results of a SELECT statement invoked as dynamic SQL. 1 Declare a dynamic cursor for the SELECT in the form: DECLARE cursor_name CURSOR FOR sql_statement_name 2 Declare the SQLDA, preferably using an INCLUDE SQLDA statement. 3 Build and PREPARE the SELECT statement. 4 Issue a DESCRIBE statement in the form: DESCRIBE sql_statement_name INTO SQLDA DESCRIBE performs the following actions: a Interrogate the database for the demographics of the expected results. b Write the addresses of the target variables to receive those results to the SQLDA. This step is bypassed if any of the following occurs: • The request does not return any data. • An INTO clause was present in the PREPARE statement. • The statement returns known columns and the INTO clause is used on the corresponding FETCH statement. • The application code defines the SQLDA. 5 Allocate storage for target variables to receive the returned data based on the demographics reported by DESCRIBE. 6 Retrieve the result rows using the following SQL cursor control statements: • OPEN cursor_name • FETCH cursor_name USING DESCRIPTOR SQLDA • CLOSE cursor_name Note that in step 6, results tables are examined one row at a time using the selection cursor. This is because client programming languages do not support data in terms of sets, but only as individual records. THIS information … IS written to this field of SQLDA … number of values to be returned SQLN column name or label of n th value SQLVAR (n th row in the SQLVAR(n) array) column data type of n th value column length of n th valueChapter 4: SQL Data Handling Event Processing Using Queue Tables SQL Reference: Fundamentals 133 Event Processing Using Queue Tables Introduction Teradata Database provides queue tables that you can use for event processing. Queue tables are base tables with first-in-first-out (FIFO) queue properties. When you create a queue table, you define a timestamp column. You can query the queue table to retrieve data from the row with the oldest timestamp. Usage An application can perform FIFO push, pop, and peek operations on queue tables. Here is an example of how an application can process events using queue tables: • Internally, you can define a trigger on a base table to insert a row into the queue table when the trigger fires. • Externally, your application can submit a SELECT AND CONSUME statement that waits for data in the queue table. • When data arrives in the queue table, the waiting SELECT AND CONSUME statement returns a result to the external application, which processes the event. Additionally, the row is deleted from the queue table. Related Topics TO perform a FIFO … USE the … push INSERT statement pop SELECT AND CONSUME statement peek SELECT statement FOR more information on … SEE … creating queue tables the CREATE/REPLACE TABLE statement in SQL Reference: Data Definition Statements SELECT AND CONSUME SQL Reference: Data Manipulation StatementsChapter 4: SQL Data Handling Manipulating Nulls 134 SQL Reference: Fundamentals Manipulating Nulls Introduction A null represents any of three things: • An empty field • An unknown value • An unknowable value Nulls are neither values nor do they signify values; they represent the absence of value. A null is a place holder indicating that no value is present. You cannot solve for the value of a null because, by definition, it has no value. For example, the expression NULL = NULL has no meaning and therefore can never be true. A query that specifies the predicate WHERE NULL = NULL is not valid because it can never be true. The meaning of the comparison it specifies is not only unknown, but unknowable. These properties make the use and interpretation of nulls in SQL problematic. The following sections outline the behavior of nulls for various SQL operations to help you to understand how to use them in data manipulation statements and to interpret the results those statements affect. NULL Literals See “NULL Keyword as a Literal” on page 90 for information on how to use the NULL keyword as a literal. Nulls and DateTime and Interval Data A DateTime or Interval value is either atomically null or it is not null. For example, you cannot have an interval of YEAR TO MONTH in which YEAR is null and MONTH is not. Result of Expressions That Contain Nulls Here are some general rules for the result of expressions that contain nulls: • When any component of a value expression is null, then the result is null. • The result of a conditional expression that has a null component is unknown. • If an operand of any arithmetic operator (such as + or -) or function (such as ABS or SQRT) is null, then the result of the operation or function is null with the exception of ZEROIFNULL. If the argument to ZEROIFNULL is NULL, then the result is 0. • COALESCE, a special shorthand variant of the CASE expression, returns NULL if all its arguments evaluate to null. Otherwise, COALESCE returns the value of the fist non-null argument. For more rules on the result of expressions containing nulls, see the sections that follow and SQL Reference: Functions and Operators.Chapter 4: SQL Data Handling Manipulating Nulls SQL Reference: Fundamentals 135 Nulls and Comparison Operators If either operand of a comparison operator is null, then the result is unknown. If either operand is the keyword NULL, an error is returned that recommends using IS NULL or IS NOT NULL instead. The following examples indicate this behavior. 5 = NULL 5 <> NULL NULL = NULL NULL <> NULL 5 = NULL + 5 Note that if the argument of the NOT operator is unknown, the result is also unknown. This translates to FALSE as a final boolean result. Instead of using comparison operators, use the IS NULL operator to search for fields that contain nulls and the IS NOT NULL operator to search for fields that do not contain nulls. For details, see “Searching for Nulls” on page 135 and “Excluding Nulls” on page 135. Using IS NULL is different from using the comparison operator =. When you use an operator like =, you specify a comparison between values or value expressions, whereas when you use the IS NULL operator, you specify an existence condition. Nulls and CASE Expressions The following rules apply to nulls and CASE expressions: • CASE and its related expressions COALESCE and NULLIF can return a null. • NULL and null expressions are valid as the CASE test expression in a valued CASE expression. • When testing for NULL, it is best to use a searched CASE expression using the IS NULL or IS NOT NULL operators in the WHEN clause. • NULL and null expressions are valid as THEN clause conditions. For details on the rules for nulls in CASE, NULLIF, and COALESCE expressions, see SQL Reference: Functions and Operators. Excluding Nulls To exclude nulls from the results of a query, use the operator IS NOT NULL. For example, to search for the names of all employees with a value other than null in the jobtitle column, enter the statement. SELECT name FROM employee WHERE jobtitle IS NOT NULL ; Searching for Nulls To search for columns that contain nulls, use the operator IS NULL. The IS NULL operator tests row data for the presence of nulls. Chapter 4: SQL Data Handling Manipulating Nulls 136 SQL Reference: Fundamentals For example, to search for the names of all employees who have a null in the deptno column, you could enter the statement: SELECT name FROM employee WHERE deptno IS NULL ; This query produces the names of all employees with a null in the deptno field. Searching for Nulls and Non-Nulls Together To search for nulls and non-nulls in the same statement, the search condition for nulls must be separate from any other search conditions. For example, to select the names of all employees with the job title of Vice Pres, Manager, or null, enter the following SELECT statement. SELECT name, jobtitle FROM employee WHERE jobtitle IN ('Manager', 'Vice Pres') OR jobtitle IS NULL ; Including NULL in the IN list has no effect because NULL never equals NULL or any value. Null Sorts as the Lowest Value in a Collation When you use an ORDER BY clause to sort records, Teradata Database sorts null as the lowest value. Sorting nulls can vary from RDBMS to RDBMS. Other systems may sort null as the highest value. If any row has a null in the column being grouped, then all rows having a null are placed into one group. NULL and Unique Indexes For unique indexes, Teradata Database treats nulls as if they are equal rather than unknown (and therefore false). For single-column unique indexes, only one row may have null for the index value; otherwise a uniqueness violation error occurs. For multi-column unique indexes, no two rows can have nulls in the same columns of the index and also have non-null values that are equal in the other columns of the index. For example, consider a two-column index. Rows can occur with the following index values: An attempt to insert a row that matches any of these rows will result in a uniqueness violation. Value of First Column in Index Value of Second Column in Index 1 null null 1 null nullChapter 4: SQL Data Handling Manipulating Nulls SQL Reference: Fundamentals 137 Teradata Database Replaces Nulls With Values on Return to Client in Record Mode When the Teradata Database returns information to a client system in record mode, nulls must be replaced with some value for the underlying column because client system languages do not recognize nulls. The following table shows the values returned for various column data types. The substitute values returned for nulls are not, by themselves, distinguishable from valid non-null values. Data from CLI is normally accessed in IndicData mode, in which additional identifying information that flags nulls is returned to the client. BTEQ uses the identifying information, for example, to determine whether the values it receives are values or just aliases for nulls so it can properly report the results. Note that BTEQ displays nulls as ?, which are not by themselves distinguishable from a CHAR or VARCHAR value of '?'. Nulls and Aggregate Functions With the important exception of COUNT(*), aggregate functions ignore nulls in their arguments. This treatment of nulls is very different from the way arithmetic operators and functions treat them. This behavior can result in apparent nontransitive anomalies. For example, if there are nulls in either column A or column B (or both), then the following expression is virtually always true. SUM(A) + (SUM B) <> SUM (A+B) Data Type Substitute Value Returned for Null CHARACTER(n) DATE (ANSI) TIME TIMESTAMP INTERVAL Pad character (or n pad characters for CHARACTER(n), where n > 1) BYTE[(n)] Binary zero byte if n omitted else n binary zero bytes VARBYTE(n) 0-length byte string VARCHARACTER(n) 0-length character string DATE (Teradata) 0 BIGINT INTEGER SMALLINT BYTEINT FLOAT DECIMAL REAL DOUBLE PRECISION NUMERIC 0Chapter 4: SQL Data Handling Session Parameters 138 SQL Reference: Fundamentals In other words, for the case of SUM, the result is never a simple iterated addition if there are nulls in the data being summed. The only exception to this is the case in which the values for columns A and B are both null in the same rows, because in those cases the entire row is disregarded in the aggregation. This is a trivial case that does not violate the general rule. The same is true, the necessary changes being made, for all the aggregate functions except COUNT(*). If this property of nulls presents a problem, you can always do either of the following workarounds, each of which produces the desired result of the aggregate computation SUM(A) + SUM(B) = SUM(A+B). • Always define NUMERIC columns as NOT NULL DEFAULT 0. • Use the ZEROIFNULL function within the aggregate function to convert any nulls to zeros for the computation, for example SUM(ZEROIFNULL(x) + ZEROIFNULL(y)) which produces the same result as this: SUM(ZEROIFNULL(x) + ZEROIFNULL(y)). COUNT(*) does include nulls in its result. For details, see SQL Reference: Functions and Operators. RANGE_N and CASE_N Functions Nulls have special considerations in the RANGE_N and CASE_N functions. For details, see SQL Reference: Functions and Operators. Session Parameters Introduction The following session parameters can be controlled with keywords or predefined system variables. Parameter Valid Keywords or System Variables SQL Flagger ON OFF Transaction Mode ANSI (COMMIT) Teradata (BTET)Chapter 4: SQL Data Handling Session Parameters SQL Reference: Fundamentals 139 SQL Flagger When enabled, the SQL Flagger assists SQL programmers by notifying them of the use of nonANSI and non-entry level ANSI SQL syntax. Enabling the SQL Flagger can be done regardless of whether you are in ANSI or Teradata session mode. Session Collation ASCII EBCDIC MULTINATIONAL HOST CHARSET_COLL JIS_COLL Account and Priority Account and reprioritization. Within the account identifier, you can specify a performance group or use one of the following predefined performance groups: • $R • $H • $M • $L Date Form ANSIDATE INTEGERDATE Character Set Indicates the character set being used by the client. You can view site-installed client character sets from DBC.CharSets or DBC.CharTranslations. The following client character sets are permanently enabled: • ASCII • EBCDIC • UTF8 • UTF16 For more information on character sets, see International Character Set Support. Express Logon (for networkattached clients) ENABLE DISABLE Parameter Valid Keywords or System VariablesChapter 4: SQL Data Handling Session Parameters 140 SQL Reference: Fundamentals To set the SQL Flagger on or off for interactive SQL, use the .SET SESSION command in BTEQ. For more detail on using the SQL Flagger, see “SQL Flagger” on page 217. To set the SQL Flagger on or off for embedded SQL, use the SQLCHECK or -sc and SQLFLAGGER or -sf options when you invoke the preprocessor. If you are using SQL in other application programs, see the reference manual for that application for instructions on enabling the SQL Flagger. Transaction Mode You can run transactions in either Teradata or ANSI session modes and these modes can be set or changed. To set the transaction mode, use the .SET SESSION command in BTEQ. For more detail on transaction semantics, see “Transaction Processing” in SQL Reference: Statement and Transaction Processing. If you are using SQL in other application programs, see the reference manual for that application for instructions on setting or changing the transaction mode. Session Collation Collation of character data is an important and complex option. The Teradata Database provides several named collations. The MULTINATIONAL and CHARSET_COLL collations allow the system administrator to provide collation sequences tailored to the needs of the site. The collation for the session is determined at logon from the defined default collation for the user. You can change your collation any number of times during the session using the SET SESSION COLLATION statement, but you cannot change your default logon in this way. Your default collation is assigned via the COLLATION option of the CREATE USER or MODIFY USER statement. This has no effect on any current session, only new logons. To set this level of flagging … Set the flag variable to this value … None SQLFLAG NONE Entry level SQLFLAG ENTRY Intermediate level SQLFLAG INTERMEDIATE To run transactions in this mode … Set the variable to this value … Teradata TRANSACTION BTET ANSI TRANSACTION ANSIChapter 4: SQL Data Handling Session Parameters SQL Reference: Fundamentals 141 Each named collation can be CASESPECIFIC or NOT CASESPECIFIC. NOT CASESPECIFIC collates lowercase data as if it were converted to uppercase before the named collation is applied. For details, see “SET SESSION COLLATION” in SQL Reference: Data Definition Statements. Account and Priority You can dynamically downgrade or upgrade the performance group priority for your account. Collation Name Description ASCII Character data is collated in the order it would appear if converted for an ASCII session, and a binary sort performed. EBCDIC Character data is collated in the order it would appear if converted for an EBCDIC session, and a binary sort performed. MULTINATIONAL The default MULTINATIONAL collation is a two-level collation based on the Unicode collation standard. Your system administrator can redefine this collation to any two-level collation of characters in the LATIN repertoire. For backward compatibility, the following are true: • MULTINATIONAL collation of KANJI1 data is single level. • The system administrator can redefine single byte character collation. This definition is not compatible with MULTINATIONAL collation of nonKANJI1 data. CHARSET_COLL collation is usually a better solution for KANJI1 data. See “ORDER BY Clause” in SQL Reference: Data Manipulation Statements. For information on setting up the MULTINATIONAL collation sequence, see “Collation Sequences” in International Character Set Support. HOST The default. HOST collation defaults are as follows: • EBCDIC collation for channel-connected systems. • ASCII collation for all others. CHARSET_COLL Character data is collated in the order it would appear if converted to the current client character set and then sorted in binary order. CHARSET_COLL collation is a system administrator-defined collation. JIS_COLL Character data is collated based on the Japanese Industrial Standards (JIS). JIS characters collate in the following order: 1 JIS X 0201-defined characters in standard order 2 JIS X 0208-defined characters in standard order 3 JIS X 0212-defined characters in standard order 4 KanjiEBCDIC-defined characters not defined in JIS X 0201, JIS X 0208, or JIS X 0212 in standard order 5 All remaining characters in Unicode standard orderChapter 4: SQL Data Handling Session Parameters 142 SQL Reference: Fundamentals Priorities can be downgraded or upgraded at either the session or the request level. For more information, see “SET SESSION ACCOUNT” in SQL Reference: Data Definition Statements. Note that changing the performance group for your account changes the account name for accounting purposes because a performance group is part of an account name. Date Form You can change the format in which DATE data is imported or exported in your current session. DATE data can be set to be treated either using the ANSI date format (DATEFORM=ANSIDATE) or using the Teradata date format (DATEFORM=INTEGERDATE). For details, see “SET SESSION DATEFORM” in SQL Reference: Data Definition Statements. Character Set To set the client character set, use one of the following: • From BTEQ, use the BTEQ [.] SET SESSION CHARSET ‘name’ command. • In a CLIv2 application, call CHARSET name. • In the URL for selecting a Teradata JDBC driver connection to a Teradata Database, use the CHARSET=name database connection parameter. where the ‘name’ or name value is ASCII, EBCDIC, UTF8, UTF16, or a name assigned to the translation codes that define an available character set. If not explicitly requested, the session default is the character set associated with the logon client. This is either the standard client default, or the character set assigned to the client by the database administrator. Express Logon Express Logon improves the logon response time for network-attached, NCR UNIX MP-RAS clients and is especially useful in the OLTP environment where sessions are short-lived. Express Logon allows the gateway to choose the fast path when logging users onto the Teradata Database. Enable or disable this mode from the Gateway Global Utility, from the XGTWGLOBAL interface: In this mode … Use this command to enable or disable Express Logon … Terminal ENABLE EXLOGON DISABLE EXLOGON Window EXLOGON button (via the LOGON dialog box)Chapter 4: SQL Data Handling Session Management SQL Reference: Fundamentals 143 The feature can be enabled or disabled for a particular host group, or for all host groups. For details on this feature, see the Utilities book. For channel-attached clients, see “Session Pools” on page 143. HELP SESSION The HELP SESSION statement identifies the transaction mode, character set, collation sequence, and date form in effect for the current session. See “HELP SESSION” in SQL Reference: Data Definition Statements for details. Session Management Introduction Each session is logged on and off via calls to CLIv2 routines or through ODBC or JDBC, which offer a one-step logon-connect function. Sessions are internally managed by dividing the session control functions into a series of single small steps that are executed in sequence to implement multi-threaded tasking. This provides concurrent processing of multiple logon and logoff events, which can be any combination of individual users, and one or more concurrent sessions established by one or more users and applications. Once connected and active, a session can be viewed as a work stream consisting of a series of requests between the client and server. Session Pools For channel-connected applications, you can establish session pools, which are collections of sessions that are logged on to the Teradata Database in advance (generally at the time of TDP initialization) for use by applications that require a ‘fast path’ logon. This capability is particularly advantageous for transaction processing in which interaction with the Teradata Database consists of many single, short transactions. TDP identifies each session with a unique session number. Teradata Database identifies a session with a session number, the username of the initiating user, and the logical host identification number of the connection (LAN or mainframe channel) associated with the controlling TDP or mTDP. For network-connected, UNIX MP-RAS applications that require fast path logons, use the Express Logon feature. For details, see “Express Logon” on page 142. Session Reserve Use the ENABLE SESSION RESERVE command from an OS/390 or VM client to reserve session capacity in the event of a PE failure. To release reserved session capacity, use the DISABLE SESSION RESERVE command. Chapter 4: SQL Data Handling Return Codes 144 SQL Reference: Fundamentals See Teradata Tools and Utilities Installation Guide for IBM OS/390 and z/OS and Teradata Tools and Utilities Installation Guide for IBM VM for further information. Session Control The major functions of session control are session logon and logoff. Upon receiving a session request, the logon function verifies authorization and returns a yes or no response to the client. The logoff function terminates any ongoing activity and deletes the session context. Requests and Responses Requests are sent to a server to initiate an action. Responses are sent by a server to reflect the results of that action. Both requests and responses are associated with an established session. A request consists of the following components: • One or more Teradata SQL statements • Control information • Optional USING data If any operation specified by an initiating request fails, the request is backed out, along with any change that was made to the database. In this case, a failure response is returned to the application. Return Codes Introduction SQL return codes provide information about the status of a completed executable SQL DML statement. Status Variables for Receiving SQL Return Codes ANSI SQL defines two status variables for receiving return codes: • SQLSTATE • SQLCODE SQLCODE is not ANSI SQL-compliant. The ANSI SQL-92 standard explicitly deprecates SQLCODE, and the ANSI SQL-99 standard does not define SQLCODE. The ANSI SQL committee recommends that new applications use SQLSTATE in place of SQLCODE. Teradata Database defines a third status variable for receiving the number of rows affected by an SQL statement in a stored procedure: • ACTIVITY_COUNT Teradata SQL defines a non-ANSI SQL Communications Area (SQLCA) that also has a field named SQLCODE for receiving return codes. Chapter 4: SQL Data Handling Return Codes SQL Reference: Fundamentals 145 Exception and Completion Conditions ANSI SQL defines two categories of conditions that issue return codes: • Exception conditions • Completion conditions Exception Conditions An exception condition indicates a statement failure. A statement that raises an exception condition does nothing more than return that exception condition to the application. There are as many exception condition return codes as there are specific exception conditions. For more information about exception conditions, see “Failure Response” on page 150 and “Error Response (ANSI Session Mode Only)” on page 149. For a complete list of exception condition codes, see the Messages book. Completion Conditions A completion condition indicates statement success. There are three categories of completion conditions: • Successful completion • Warnings • No data found For more information, see: • “Statement Responses” on page 147 • “Success Response” on page 148 • “Warning Response” on page 149 A statement that raises a completion condition can take further action such as querying the database and returning results to the requesting application, updating the database, initiating an SQL transaction, and so on. For information on … See … • SQLSTATE • SQLCODE • ACTIVITY_COUNT “Result Code Variables” in SQL Reference: Stored Procedures and Embedded SQL SQLCA “SQL Communications Area (SQLCA)” in SQL Reference: Stored Procedures and Embedded SQLChapter 4: SQL Data Handling Return Codes 146 SQL Reference: Fundamentals Return Codes for Stored Procedures The return code values are different in the case of SQL control statements in stored procedures. The return codes for stored procedures appear in the following table. How an Application Uses SQL Return Codes An application program or stored procedure tests the status of a completed executable SQL statement to determine its status. FOR this type of completion condition … The value for this return code is … SQLSTATE SQLCODE Success '00000' 0 Warning '01901' 901 '01800' to '01841' 901 '01004' 902 No data found '02000' 100 FOR this type of condition … The value for this return code is … SQLSTATE SQLCODE Successful completion '00000' 0 Warning SQLSTATE value corresponding to the warning code. the Teradata Database warning code. No data found or any other Exception SQLSTATE value corresponding to the error code. the Teradata Database error code. IF the statement raises this type of condition … THEN the application or condition handler takes the following remedial action … Successful completion none. Warning the statement execution continues. If a warning condition handler is defined in the application, the handler executes.Chapter 4: SQL Data Handling Statement Responses SQL Reference: Fundamentals 147 Statement Responses Response Types The Teradata Database responds to an SQL request with one of the following condition responses: • Success response, with optional warning • Failure response • Error response (ANSI session mode only) Depending on the type of statement, the Teradata Database also responds with one or more rows of data. Multistatement Responses A response to a request that contains more than one statement, such as a macro, is not returned to the client until all statements in the request are successfully executed. How a Response Is Returned to the User The manner in which the response is returned depends on the interface that is being used. For example, if an application is using a language preprocessor, then the activity count, warning code, error code, and fields from a selected row are returned directly to the program through its appropriately declared variables. If the application is a stored procedure, then the activity count is returned directly in the ACTIVITY_COUNT status variable. If you are using BTEQ, then a success, error, or failure response is displayed automatically. Response Condition Codes SQL statements also return condition codes that are useful for handling errors and warnings in embedded SQL and stored procedure applications. No data found or any other exception whatever appropriate action is required by the exception. If an EXIT handler has been defined for the exception, the statement execution terminates. If a CONTINUE handler has been defined, execution continues after the remedial action. IF the statement raises this type of condition … THEN the application or condition handler takes the following remedial action …Chapter 4: SQL Data Handling Success Response 148 SQL Reference: Fundamentals For information about SQL response condition codes, see the following in SQL Reference: Stored Procedures and Embedded SQL: • SQLSTATE • SQLCODE • ACTIVITY_COUNT Success Response Definition A success response contains an activity count that indicates the total number of rows involved in the result. For example, the activity count for a SELECT statement is the total number of rows selected for the response. For a SELECT, COMMENT, or ECHO statement, the activity count is followed by the data that completes the response. An activity count is meaningful for statements that return a result set, for example: • SELECT • INSERT • UPDATE • DELETE • HELP • SHOW • EXPLAIN • CREATE PROCEDURE • REPLACE PROCEDURE For other SQL statements, activity count is meaningless. Example The following interactive SELECT statement returns the successful response message. SELECT AVG(f1) FROM Inventory; *** Query completed. One row found. One column returned. *** Total elapsed time was 1 second. Average(f1) ----------- 14Chapter 4: SQL Data Handling Warning Response SQL Reference: Fundamentals 149 Warning Response Definition A success or OK response with a warning indicates either that an anomaly has occurred or informs the user about the anomaly and indicates how it can be important to the interpretation of the results returned. Example Assume the current session is running in ANSI session mode. If nulls are included in the data for column f1, then the following interactive query returns the successful response message with a warning about the nulls. SELECT AVG(f1) FROM Inventory; *** Query completed. One row found. One column returned. *** Warning: 2892 Null value eliminated in set function. *** Total elapsed time was 1 second. Average(f1) ----------- 14 This warning response is not generated if the session is running in Teradata session mode. Error Response (ANSI Session Mode Only) Definition An error response occurs when a query anomaly is severe enough to prevent the correct processing of the request. In ANSI session mode, an error for a request causes the request to rollback, and not the entire transaction. Example 1 The following command returns the error message immediately following. .SET SESSION TRANS ANSI; *** Error: You must not be logged on .logoff to change the SQLFLAG or TRANSACTION settings. Example 2 Assume that the session is running in ANSI session mode, and the following table is defined: CREATE MULTISET TABLE inv, FALLBACK, NO BEFORE JOURNAL, NO AFTER JOURNALChapter 4: SQL Data Handling Failure Response 150 SQL Reference: Fundamentals ( item INTEGER CHECK ((item >=10) AND (item <= 20) )) PRIMARY INDEX (item); You insert a value of 12 into the item column of the inv table. This is valid because the defined integer check specifies that any integer between 10 and 20 (inclusive) is valid. INSERT INTO inv (12); The following results message returns. *** Insert completed. One row added.... You insert a value of 9 into the item column of the inv table. This is not valid because the defined integer check specifies that any integer with a value less than 10 is not valid. INSERT INTO inv (9); The following error response returns: ***Error 5317 Check constraint violation: Check error in field inv.item. You commit the current transaction: COMMIT; The following results message returns: *** COMMIT done. ... You select all rows from the inv table: SELECT * FROM inv; The following results message returns: *** Query completed. One row found. One column returned. item ------- 12 Failure Response Definition A failure response is a severe error. The response includes a statement number, an error code, and an associated text string describing the cause of the failure. Teradata Session Mode In Teradata session mode, a failure causes the system to roll back the entire transaction. If one statement in a macro fails, a single failure response is returned to the client, and the results of any previous statements in the transaction are backed out.Chapter 4: SQL Data Handling Failure Response SQL Reference: Fundamentals 151 ANSI Session Mode In ANSI session mode, a failure causes the system to roll back the entire transaction, for example, when the current request: • Results in a deadlock • Performs a DDL statement that aborts • Executes an explicit ROLLBACK or ABORT statement Example 1 The following SELECT statement SELECT * FROM Inventory:; in BTEQ, returns the failure response message: *** Failure 3709 Syntax error, replace the ':' that follows the name with a ';'. Statement# 1, Info =20 *** Total elapsed time was 1 second. Example 2 Assume that the session is running in ANSI session mode, and the following table is defined: CREATE MULTISET TABLE inv, FALLBACK, NO BEFORE JOURNAL, NO AFTER JOURNAL ( item INTEGER CHECK ((item >=10) AND (item <= 20) )) PRIMARY INDEX (item); You insert a value of 12 into the item column of the inv table. This is valid because the defined integer check specifies that any integer between 10 and 20 (inclusive) is valid. INSERT INTO inv (12); The following results message returns. *** Insert completed. One row added.... You commit the current transaction: COMMIT; The following results message returns: *** COMMIT done. ... You insert a valid value of 15 info the item column of the inv table: INSERT INTO inv (15); The following results message returns. *** Insert completed. One row added....Chapter 4: SQL Data Handling Failure Response 152 SQL Reference: Fundamentals You can use the ABORT statement to cause the system to roll back the transaction: ABORT; The following failure message returns: *** Failure 3514 User-generated transaction ABORT. Statement# 1, Info =0 You select all rows from the inv table: SELECT * FROM inv; The following results message returns: *** Query completed. One row found. One column returned. item ------- 12SQL Reference: Fundamentals 153 CHAPTER 5 Query Processing This chapter discusses query processing, including single AMP requests and all AMP requests, and table access methods available to the Optimizer. Topics include: • Query processing • Table access methods • Full-table scans • Collecting statistics Query Processing Introduction An SQL query (the definition for “query” here includes DELETE, INSERT, MERGE, and UPDATE as well as SELECT) can affect one AMP, several AMPs, or all AMPs in the configuration. IF a query … THEN … involving a single table uses a unique primary index (UPI) the row hash can be used to identify a single AMP. At most one row can be returned. involving a single table uses a nonunique primary index (NUPI) the row hash can be used to identify a single AMP. Any number of rows can be returned. uses a unique secondary index (USI) one or two AMPs are affected (one AMP if the subtable and base table are on the same AMP). At most one row can be returned. uses a nonunique secondary index (NUSI) if the table has a partitioned primary index (PPI) and the NUSI is the same column set as a NUPI, the query affects one AMP. Otherwise, all AMPs take part in the operation and any number of rows can be returned.Chapter 5: Query Processing Query Processing 154 SQL Reference: Fundamentals The SELECT statements in subsequent examples reference the following table data. Single AMP Request Assume that a PE receives the following SELECT statement: SELECT last_name FROM Employee WHERE employee_number = 1008; Because a unique primary index value is used as the search condition (the column employee_number is the primary index for the Employee table), PE1 generates a single AMP step requesting the row for employee 1008. The AMP step, along with the PE identification, is put into a message, and sent via the BYNET to the relevant AMP (processor). This process is illustrated by the graphic under “Flow Diagram of a Single AMP Request” on page 155. Only one BYNET is shown to simplify the illustration. Abbreviation Meaning PK Primary Key FK Foreign Key UPI Unique Primary Index Employee Employee Number Manager Employee Number Dept. Number Job Code Last Name First Name Hire Date Birth Date Salary Amount PK/UPI FK FK FK 1006 1019 301 312101 Stein John 76105 531015 2945000 1008 1019 301 312102 Kanieski Carol 770201 580517 2925000 1005 0801 403 431100 Ryan Loretta 761015 550910 3120000 1004 1003 401 412101 Johnson Darlene 761015 460423 3630000 1007 1005 403 432101 Villegas Arnando 770102 370131 4970000 1003 0801 401 411100 Trader James 760731 470619 3755000 1016 0801 302 321100 Rogers Nora 780310 590904 5650000 1012 1005 403 432101 Hopkins Paulene 770315 420218 3790000 1019 0801 301 311100 Kubic Ron 780801 421211 5770000 1023 1017 501 512101 Rabbit Peter 790301 621029 2650000 1083 0801 619 414221 Kimble George 910312 410330 3620000 1017 0801 501 511100 Runyon Irene 780501 511110 6600000 1001 1003 401 412101 Hoover William 760818 500114 2552500Chapter 5: Query Processing Query Processing SQL Reference: Fundamentals 155 Flow Diagram of a Single AMP Request Assuming that AMP2 has the row, it accepts the message. As illustrated by the graphic under “Single AMP Response to Requesting PE” on page 156, AMP2 retrieves the row from its DSU (disk storage unit), includes the row and the PE identification in a return message, and sends the message back to PE1 via the BYNET. PE1 accepts the message and returns the response row to the requesting application. For an illustration of a single AMP request with partition elimination, see “Single AMP Request With Partition Elimination” on page 160. BYNET PE1 PE2 AMP1 AMP2 AMP3 AMP4 1006 STEIN 1008 KANIESKI 1023 RABBIT 1004 JOHNSON 1101C002 AMP STEP DSU DSU DSUChapter 5: Query Processing Query Processing 156 SQL Reference: Fundamentals Single AMP Response to Requesting PE All AMP Request Assume PE1 receives a SELECT statement that specifies a range of primary index values as a search condition as shown in the following example: SELECT last_name, employee_number FROM employee WHERE employee_number BETWEEEN 1001 AND 1010 ORDER BY last_name; In this case, each value hashes differently, and all AMPs must search for the qualifying rows. PE1 first parses the request and creates the following AMP steps: • Retrieve rows between 1001 and 1010 • Sort ascending on last_name • Merge the sorted rows to form the answer set PE1 then builds a message for each AMP step and puts that message onto the BYNET. Typically, each AMP step is completed before the next one begins; note, however, that some queries can generate parallel steps. When PE1 puts the message for the first AMP step on the BYNET, that message is broadcast to all processors as illustrated by “Figure 1: Flow Diagram for an All AMP Request” on page 157. BYNET ROW 1008 PE1 PE2 AMP1 AMP2 AMP3 AMP4 1006 Stein 1008 Kanieski 1023 Rabbit 1004 Johnson 1101C003Chapter 5: Query Processing Query Processing SQL Reference: Fundamentals 157 Figure 1: Flow Diagram for an All AMP Request The process is as follows: 1 All AMPs accept the message, but the PEs do not. 2 Each AMP checks for qualifying rows on its disk storage units. 3 If any qualifying rows are found, the data in the requested columns is converted to the client format and copied to a spool file. 4 Each AMP completes the step, whether rows were found or not, and puts a completion message on the BYNET. The completion messages flow across the BYNET to PE1. 5 When all AMPs have returned a completion message, PE1 transmits a message containing AMP Step 2 to the BYNET. Upon receipt of Step 2, the AMPs sort their individual answer sets into ascending sequence by last_name (see “Figure 2: Flow Diagram for an AMP Sort” on page 158). Note: If partitioned on employee_number, the scan may be limited to a few partitions based on partition elimination. PE1 PE2 AMP1 AMP2 AMP3 AMP4 1006 STEIN 1008 KANIESKI 1004 JOHNSON 1007 VILLEGAS 1003 TRADER 1001 HOOVER 1005 RYAN BYNET FF02A004 DATA SPOOLChapter 5: Query Processing Query Processing 158 SQL Reference: Fundamentals Figure 2: Flow Diagram for an AMP Sort 6 Each AMP sorts its answer set, then puts a completion message on the BYNET. 7 When PE1 has received all completion messages for Step 2, it sends a message containing AMP Step 3. 8 Upon receipt of Step 3, each AMP copies the first block from its sorted spool to the BYNET. Because there can be multiple AMPs on a single node, each node might be required to handle sort spools from multiple AMPs (see “Figure 3: Flow Diagram for a BYNET Merge” on page 159). PE1 PE2 AMP1 AMP2 AMP3 AMP4 1006 STEIN 1008 KANIESKI 1004 JOHNSON 1007 VILLEGAS 1003 TRADER 1001 HOOVER 1005 RYAN BYNET FF02A005 1004 JOHNSON 1008 KANIESKI 1006 STEIN 1003 TRADER 1007 VILLEGAS 1001 HOOVER 1005 RYAN DATA SPOOL SORT SPOOLChapter 5: Query Processing Query Processing SQL Reference: Fundamentals 159 Figure 3: Flow Diagram for a BYNET Merge 9 Nodes that contain multiple AMPs must first perform an intermediate sort of the spools generated by each of the local AMPs. When the local sort is complete on each node, the lowest sorting row from each node is sent over the BYNET to PE1. From this point on, PE1 acts as the Merge coordinator among all the participating nodes. 10 The Merge continues with PE1 building a globally sorted buffer. When this buffer fills, PE1 forwards it to the application and begins building subsequent buffers. 11 When a participant node has exhausted its sort spool, it sends a Done message to PE1. This causes PE1 to prune this node from the set of Merge participants. When there are no remaining Merge participants, PE1 sends the final buffer to the application along with an End Of File message. Partition Elimination A PPI can increase query efficiency via partition elimination. The degree of partition elimination depends on the: • Partition expression for the primary index of the table • Conditions in the query • Capability of the Optimizer to detect partition elimination It is not always required that all values of the partitioning columns be specified in a query to have partition elimination occur. HD03A005 Global Sort Buffer Local Sort Tree Sort Spools Node 1 Local Sort Tree Sort Spools Node 3 Sort Spools Node 2 Local Sort Tree Local Sort Tree Sort Spools Node 4 PE1 PE2 AMP AMP AMP AMP PE5 PE6 AMP AMP AMP AMP PE7 PE8 AMP AMP AMP AMP PE3 PE4 AMP AMP AMP AMP BYNETChapter 5: Query Processing Query Processing 160 SQL Reference: Fundamentals Single AMP Request With Partition Elimination If a SELECT specifies values for all the primary index columns, the AMP where the rows reside can be determined and only a single AMP is accessed. If conditions are also specified on the partitioning columns, partition elimination may reduce the number of partitions to be probed on that AMP. IF a SELECT … THEN … specifies values for all the primary index columns the AMP where the rows reside can be determined and only a single AMP is accessed. IF conditions are … THEN … not specified on the partitioning columns each partition can be probed to find the rows based on the hash value. also specified on the partitioning columns partition elimination may reduce the number of partitions to be probed on that AMP. For an illustration, see “Single AMP Request With Partition Elimination” on page 160. does not specify the values for all the primary index columns an all-AMP full file scan is required for a table with an NPPI. However, with a PPI, if conditions are specified on the partitioning columns, partition elimination may reduce an all-AMP full file scan to an all-AMP scan of only the non-eliminated partitions.Chapter 5: Query Processing Table Access SQL Reference: Fundamentals 161 The following diagram illustrates this process. The AMP Step includes the list of partitions (P) to access. Partition elimination reduces access to the partitions that satisfy the query requirements. In each partition, look for rows with a given row hash value (RH) of the PI. Table Access Teradata Database uses indexes and partitions to access the rows of a table. If indexed or partitioned access is not suitable for a query, the result is a full-table scan. Access Methods The following table access methods are available to the Optimizer: BYNET PE1 PE2 AMP1 AMP2 AMP3 AMP4 Table P P 1101A094 AMP STEP DSU DSU DSU RH RH RH RH • Unique Primary Index • Unique Partitioned Primary Index • Nonunique Primary Index • Nonunique Partitioned Primary Index • Unique Secondary Index • Nonunique Secondary Index • Join Index • Hash Index • Full-Table Scan • Partition ScanChapter 5: Query Processing Table Access 162 SQL Reference: Fundamentals Effects of Conditions in WHERE Clause Whether the system can use row hashing, or do a table scan with partition elimination, or whether it must do a full-table scan depends on the predicates or conditions that appear in the WHERE clause associated with an UPDATE, DELETE, or SELECT statement. The following functions are applied to rows identified by the WHERE clause, and have no effect on the selection of rows from the base table: Statements that specify any of the following WHERE clause conditions result in full-table scans (FTS). If the table has a PPI, partition elimination might reduce the FTS access to only the affected partitions. The type of table access that the system uses when statements specify any of the following WHERE clause conditions depends on whether the column or columns are indexed, the type of index, and its selectivity: • GROUP BY • HAVING • INTERSECT • MINUS/EXCEPT • ORDER BY • QUALIFY • SAMPLE • UNION • WITH ... BY • WITH • nonequality comparisons • column_name IS NOT NULL • column_name NOT IN (explicit list of values) • column_name NOT IN (subquery) • column_name BETWEEN ... AND ... • condition_1 OR condition_2 • NOT condition_1 • column_name LIKE • column_1 || column_2 = value • table1.column_x = table1.column_y • table1.column_x [computation] = value • table1.column_x [computation] - table1.column_y • INDEX (column_name) • SUBSTR (column_name) • SUM • MIN • MAX • AVG • DISTINCT • COUNT • ANY • ALL • missing WHERE clause • column_name = value or constant expression • column_name IS NULL • column_name IN (explicit list of values) • column_name IN (subquery) • condition_1 AND condition_2 • different data types • table1.column_x = table2.column_xChapter 5: Query Processing Full-Table Scans SQL Reference: Fundamentals 163 In summary, a query influences processing choices as follows: • A full-table scan (possibly with partition elimination if the table has a PPI) is required if the query includes an implicit range of values, such as in the following WHERE examples. Note that when a small BETWEEN range is specified, the optimizer can use row hashing rather than a full-table scan. ... WHERE column_name [BETWEEN <, >, <>, <=, >=] ... WHERE column_name [NOT] IN (SELECT...) ... WHERE column_name NOT IN (val1, val2 [,val3]) • Row hashing can be used if the query includes an explicit value, as shown in the following WHERE examples: ... WHERE column_name = val ... WHERE column_name IN (val1, val2, [,val3]) Related Topics Full-Table Scans Introduction A full-table scan is a retrieval mechanism that touches all rows in a table. If you do not specify a WHERE clause in your query, then the Teradata Database always uses a full-table scan to access the data. Even when results are qualified using a WHERE clause, indexed or partitioned access may not be suitable for a query, and a full-table scan may result. A full-table scan is always an all-AMP operation, and should be avoided when possible. Fulltable scans may generate spool files that can have as many rows as the base table. Full-table scans are not something to fear, however. The architecture that the Teradata Database uses makes a full-table scan an efficient procedure, and optimization is scalable based on the number of AMPs defined for the system. The sorts of unplanned, ad hoc queries that characterize the data warehouse process, and that often are not supported by indexes, perform very effectively for Teradata Database using full-table scans. FOR more information on … SEE … the efficiency, number of AMPs used, and the number of rows accessed by all table access methods Database Design strengths and weaknesses of table access methods Introduction to Teradata Warehouse full-table scans “Full-Table Scans” on page 163 index access “Indexes” on page 17Chapter 5: Query Processing Collecting Statistics 164 SQL Reference: Fundamentals How a Full-Table Scan Accesses Rows Because full-table scans necessarily touch every row on every AMP, they do not use the following mechanisms for locating rows. • Hashing algorithm and hash map • Primary indexes • Secondary indexes or their subtables • Partitioning Instead, a full-table scan uses the file system tables known as the Master Index and Cylinder Index to locate each data block. Each row within a data block is located by a forward scan. Because rows from different tables are never mixed within the same data block and because rows never span blocks, an AMP can scan up to 128K bytes of the table on each block read, making a full-table scan a very efficient operation. Data block read-ahead and cylinder reads can also increase efficiency. Related Topics Collecting Statistics The COLLECT STATISTICS (Optimizer form) statement collects demographic data for one or more columns of a base table, hash index, or join index, computes a statistical profile of the collected data, and stores the synopsis in the data dictionary. The Optimizer uses the synopsis data when it generates its table access and join plans. Usage You should collect statistics on newly created, empty data tables. An empty collection defines the columns, indexes, and synoptic data structure for loaded collections. You can easily collect statistics again after the table is populated for prototyping, and again when it is in production. FOR more information on … SEE … full-table scans Database Design cylinder reads Database Administration data-block read ahead • Performance Management • DBS Control Utility in UtilitiesChapter 5: Query Processing Collecting Statistics SQL Reference: Fundamentals 165 You can collect statistics on a: • Unique index, which can be: • Primary or secondary • Single or multiple column • Partitioned or non-partitioned • Non-unique index, which can be: • Primary or secondary • Single or multiple column • Partitioned or non-partitioned • With or without COMPRESS fields • Non-indexed column or set of columns, which can be: • Partitioned or non-partitioned • With or without COMPRESS fields • Join index • Hash index • Temporary table • If you specify the TEMPORARY keyword but a materialized table does not exist, the system first materializes an instance based on the column names and indexes you specify. This means that after a true instance is created, you can update (re-collect) statistics on the columns by entering COLLECT STATISTICS and the TEMPORARY keyword without having to specify the desired columns and index. • If you omit the TEMPORARY keyword but the table is a temporary table, statistics are collected for an empty base table rather than the materialized instance. • Sample (system-selected percentage) of the rows of a data table or index, to detect data skew and dynamically increase the sample size when found. • The SAMPLE option is not supported for global temporary tables, join indexes, or hash indexes. • The system does not store both sampled and defined statistics for the same index or column set. Once sampled statistics have been collected, implicit re-collection hits the same columns and indexes, and operates in the same mode. To change this, specify any keywords or options and name the columns and/or indexes.Chapter 5: Query Processing Collecting Statistics 166 SQL Reference: Fundamentals Related Topics FOR more information on … SEE … using the COLLECT STATISTICS statement SQL Reference: Data Definition Statements collecting statistics on a join index Database Design collecting statistics on a hash index when to collect statistics on base table columns instead of hash index columns database administration and collecting statistics Database AdministrationSQL Reference: Fundamentals 167 APPENDIX A Notation Conventions This appendix describes the notation conventions used in this book. Throughout this book, three conventions are used to describe the SQL syntax and code: • Syntax diagrams, used to describe SQL syntax form, including options. See “Syntax Diagram Conventions” on page 167. • Square braces in the text, used to represent options. The indicated parentheses are required when you specify options. For example: • DECIMAL [(n[,m])] means the decimal data type can be defined optionally: • without specifying the precision value n or scale value mspecifying precision (n) only • specifying both values (n,m) • you cannot specify scale without first defining precision. • CHARACTER [(n)] means that use of (n) is optional. The values for n and m are integers in all cases • Japanese character code shorthand notation, used to represent unprintable Japanese characters. See “Character Shorthand Notation Used In This Book” on page 171. Symbols from the predicate calculus are also used occasionally to describe logical operations. See “Predicate Calculus Notation Used in This Book” on page 172. Syntax Diagram Conventions Notation Conventions The following table defines the notation used in this section: Item Definition / Comments Letter An uppercase or lowercase alphabetic character ranging from A through Z. Number A digit ranging from 0 through 9. Do not use commas when entering a number with more than three digits.Appendix A: Notation Conventions Syntax Diagram Conventions 168 SQL Reference: Fundamentals Paths The main path along the syntax diagram begins at the left, and proceeds, left to right, to the vertical bar, which marks the end of the diagram. Paths that do not have an arrow or a vertical bar only show portions of the syntax. The only part of a path that reads from right to left is a loop. Paths that are too long for one line use continuation links. Continuation links are small circles with letters indicating the beginning and end of a link: When you see a circled letter in a syntax diagram, go to the corresponding circled letter and continue. Word Variables and reserved words. IF a word is shown in … THEN it represents … UPPERCASE LETTERS a keyword. Syntax diagrams show all keywords in uppercase, unless operating system restrictions require them to be in lowercase. If a keyword is shown in uppercase, you may enter it in uppercase or mixed case. lowercase letters a keyword that you must enter in lowercase, such as a UNIX command. lowercase italic letters a variable such as a column or table name. You must substitute a proper value. lowercase bold letters a variable that is defined immediately following the diagram that contains it. UNDERLINED LETTERS the default value. This applies both to uppercase and to lowercase words. Spaces Use one space between items, such as keywords or variables. Punctuation Enter all punctuation exactly as it appears in the diagram. Item Definition / Comments FE0CA002 A AAppendix A: Notation Conventions Syntax Diagram Conventions SQL Reference: Fundamentals 169 Required Items Required items appear on the main path: If you can choose from more than one item, the choices appear vertically, in a stack. The first item appears on the main path: Optional Items Optional items appear below the main path: If choosing one of the items is optional, all the choices appear below the main path: You can choose one of the options, or you can disregard all of the options. Abbreviations If a keyword or a reserved word has a valid abbreviation, the unabbreviated form always appears on the main path. The shortest valid abbreviation appears beneath. In the above syntax, the following formats are valid: • SHOW CONTROLS • SHOW CONTROL Loops A loop is an entry or a group of entries that you can repeat one or more times. Syntax diagrams show loops as a return path above the main path, over the item or items that you can repeat. FE0CA003 SHOW FE0CA005 SHOW VERSIONS CONTROLS FE0CA004 SHOW CONTROLS FE0CA006 SHOW CONTROLS VERSIONS FE0CA042 SHOW CONTROL CONTROLSAppendix A: Notation Conventions Syntax Diagram Conventions 170 SQL Reference: Fundamentals The following rules apply to loops: Excerpts Sometimes a piece of a syntax phrase is too large to fit into the diagram. Such a phrase is indicated by a break in the path, marked by | terminators on either side of the break. A name for the excerpted piece appears between the break marks in boldface type. The named phrase appears immediately after the complete diagram, as illustrated by the following example. IF … THEN … there is a maximum number of entries allowed the number appears in a circle on the return path. In the example, you may enter cname a maximum of 4 times. there is a minimum number of entries required the number appears in a square on the return path. In the example, you must enter at least 3 groups of column names. a separator character is required between entries the character appears on the return path. If the diagram does not show a separator character, use one blank space. In the example, the separator character is a comma. a delimiter character is required around entries the beginning and end characters appear outside the return path. Generally, a space is not needed between delimiter characters and entries. In the example, the delimiter characters are the left and right parentheses. JC01B012 ( , 4 cname ) , 3 LOCKING excerpt where_cond A cname excerpt JC01A014 A HAVING con , col_pos ,Appendix A: Notation Conventions Character Shorthand Notation Used In This Book SQL Reference: Fundamentals 171 Character Shorthand Notation Used In This Book Introduction This book uses the UNICODE naming convention for characters. For example, the lowercase character ‘a’ is more formally specified as either LATIN SMALL LETTER A or U+0041. The U+xxxx notation refers to a particular code point in the Unicode standard, where xxxx stands for the hexadecimal representation of the 16-bit value defined in the standard. In parts of the book, it is convenient to use a symbol to represent a special character, or a particular class of characters. This is particularly true in discussion of the following Japanese character encodings. • KanjiEBCDIC • KanjiEUC • KanjiShift-JIS These encodings are further defined in the International Character Set Support book. Symbols The symbols, and the character sets with which they are used, are defined in the following table. Symbol Encoding Meaning a..z A..Z 0..9 Any Any single byte Latin letter or digit. a..z A..Z 0..9 Unicode compatibility zone Any fullwidth Latin letter or digit. < KanjiEBCDIC Shift Out [SO] (0x0E). Indicates transition from single to multibyte character in KanjiEBCDIC. > KanjiEBCDIC Shift In [SI] (0x0F). Indicates transition from multibyte to single byte KanjiEBCDIC. T Any Any multibyte character. Its encoding depends on the current character set. For KanjiEUC, “ss 3 ” sometimes precedes code set 3 characters. I Any Any single byte Hankaku Katakana character. In KanjiEUC, it must be preceded by “ss 2 ”, forming an individual multibyte character. ? Any Represents the graphic pad character.Appendix A: Notation Conventions Predicate Calculus Notation Used in This Book 172 SQL Reference: Fundamentals For example, string “TEST”, where each letter is intended to be a fullwidth character, is written as TEST. Occasionally, when encoding is important, hexadecimal representation is used. For example, the following mixed single byte/multibyte character data in KanjiEBCDIC character set LMNQRS is represented as: D3 D4 D5 0E 42E3 42C5 42E2 42E3 0F D8 D9 E2 Pad Characters The following table lists the pad characters for the various server character sets. Predicate Calculus Notation Used in This Book Relational databases are based on the theory of relations as developed in set theory. Predicate calculus is often the most unambiguous way to express certain relational concepts. Occasionally this book uses the following predicate calculus notation to explain concepts. ? Any Represents either a single or multibyte pad character, depending on context. ss 2 KanjiEUC Represents the EUC code set 2 introducer (0x8E). ss 3 KanjiEUC Represents the EUC code set 3 introducer (0x8F). Symbol Encoding Meaning Server Character Set Pad Character Name Pad Character Value LATIN SPACE 0x20 UNICODE SPACE U+0020 GRAPHIC IDEOGRAPHIC SPACE U+3000 KANJISJIS SPACE 0x20 KANJI1 SPACE 0x20 This symbol … Represents this phrase … iff If and only if ? For all ? There existsSQL Reference: Fundamentals 173 APPENDIX B Restricted Words for V2R6.2 This appendix details restrictions for Release V2R6.2 on the use of certain terminology in SQL queries and in other user application programs that interface with the Teradata Database. The following sections are described: • A current listing of Teradata reserved keywords, non-reserved keywords, those words reserved for future use, and ANSI SQL-2003 reserved and non-reserved keywords. • Statements about the varying usage restrictions of each type of word. Reserved Words and Keywords for V2R6.2 The following list contains all classes of restricted words for Teradata Database Release V2R6.2, and uses these conventions: • Abbreviations and the full words they represent appear separately, except in cases where the abbreviation is the only common usage, such as ASCII. • The following definitions apply to the Teradata Database Status column: Type Explanation Reserved Teradata Database reserved word that cannot be used as an identifier to name host variables, correlations, local variables in stored procedures, objects, such as databases, tables, columns, or stored procedures, or parameters, such as macro or stored procedure parameters, because Teradata Database already uses the word and might misinterpret it. Future Word reserved for future Teradata Database use and cannot be used as an identifier. NonReserved Teradata Database non-reserved keyword that is permitted as an identifier but discouraged because of possible confusion that may result. empty If the keyword does not have a Teradata Database status, the word is permitted as an identifier but discouraged because it is an SQL-2003 reserved or non-reserved word.Appendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 174 SQL Reference: Fundamentals • The following definitions apply to the SQL-2003 Status column: Type Explanation Reserved ANSI SQL-2003 reserved word. If the Teradata Database Status is Reserved or Future, an SQL-2003 reserved word cannot be used as an identifier. If the Teradata Database Status is Non-Reserved or empty, the word is permitted as an identifier but discouraged because of possible confusion that may result. NonReserved ANSI SQL-2003 non-reserved word. If the Teradata Database Status is Reserved or Future, an SQL-2003 non-reserved word cannot be used as an identifier. If the Teradata Database Status is Non-Reserved or empty, the word is permitted as an identifier, but discouraged because of the possible confusion that may result. Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReserved A X ABORT X ABORTSESSION X ABS X X ABSOLUTE X ACCESS X ACCESS_LOCK X ACCOUNT X ACOS X ACOSH X ACTION X ADA X ADD X X ADD_MONTHS X ADMIN X X AFTER X X AG X AGGREGATE XAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 175 ALIAS X ALL X X ALLOCATE X ALLOCATION X ALTER X X ALWAYS X X AMP X ANALYSIS X AND X X ANSIDATE X ANY X X ARE X ARGLPAREN X ARRAY X AS X X ASC X X ASCII X ASENSITIVE X ASIN X ASINH X ASSERTION X ASSIGNMENT X X ASYMMETRIC X AT X X ATAN X ATAN2 X ATANH X ATOMIC X X ATTR X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 176 SQL Reference: Fundamentals ATTRIBUTE X X ATTRIBUTES X X ATTRS X AUTHORIZATION X X AVE X AVERAGE X AVG X X BEFORE X X BEGIN X X BERNOULLI X BETWEEN X X BIGINT X X BINARY X X BLOB X X BOOLEAN X BOTH X X BREADTH X BT X BUT X BY X X BYTE X BYTEINT X BYTES X C X X CALL X X CALLED X X CARDINALITY X CASCADE X CASCADED X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 177 CASE X X CASE_N X CASESPECIFIC X CAST X X CATALOG X CATALOG_NAME X CD X CEIL X CEILING X CHAIN X CHANGERATE X CHAR X X CHAR_LENGTH X X CHAR2HEXINT X CHARACTER X X CHARACTER_LENGTH X X CHARACTER_SET_CATALOG X CHARACTER_SET_NAME X CHARACTER_SET_SCHEMA X CHARACTERISTICS X X CHARACTERS X X CHARS X CHARSET_COLL X CHECK X X CHECKED CHECKPOINT X CHECKSUM X CLASS X CLASS_ORIGIN X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 178 SQL Reference: Fundamentals CLOB X X CLOSE X X CLUSTER X CM X COALESCE X X COBOL X COLLATE X COLLATION X X COLLATION_CATALOG X COLLATION_NAME X COLLATION_SCHEMA X COLLECT X X COLUMN X X COLUMN_NAME X COLUMNSPERINDEX X COLUMNSPERJOININDEX X COMMAND_FUNCTION X COMMAND_FUNCTION_CODE X COMMENT X COMMIT X X COMMITTED X COMPARISON X COMPILE X COMPRESS X CONDITION X CONDITION_NUMBER X CONNECT X CONNECTION X CONNECTION_NAME X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 179 CONSTRAINT X X CONSTRAINT_CATALOG X CONSTRAINT_NAME X CONSTRAINT_SCHEMA X CONSTRAINTS X CONSTRUCTOR X X CONSUME X CONTAINS X CONTINUE X X CONVERT X CONVERT_TABLE_HEADER X CORR X X CORRESPONDING X COS X COSH X COSTS X COUNT X X COVAR_POP X X COVAR_SAMP X X CPP X CPUTIME X CREATE X X CROSS X X CS X CSUM X CT X CUBE X X CUME_DIST X CURRENT X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 180 SQL Reference: Fundamentals CURRENT_DATE X X CURRENT_DEFAULT_TRANSFORM_GROUP X CURRENT_PATH X CURRENT_ROLE X CURRENT_TIME X X CURRENT_TIMESTAMP X X CURRENT_TRANSFORM_GROUP_FOR_TYPE X CURRENT_USER X CURSOR X X CURSOR_NAME X CV X CYCLE X X DATA X X DATABASE X DATABLOCKSIZE X DATE X X DATEFORM X DATETIME_INTERVAL_CODE X DATETIME_INTERVAL_PRECISION X DAY X X DBC X DEALLOCATE X DEBUG X DEC X X DECIMAL X X DECLARE X X DEFAULT X X DEFAULTS X DEFERRABLE X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 181 DEFERRED X X DEFINED X DEFINER X X DEGREE X DEGREES X DEL X DELETE X X DEMOGRAPHICS X DENIALS X DENSE_RANK X DEPTH X DEREF X DERIVED X DESC X X DESCRIBE X DESCRIPTOR X X DETERMINISTIC X X DIAGNOSTIC X DIAGNOSTICS X DIGITS X DISABLED X DISCONNECT X DISPATCH X DISTINCT X X DO X DOMAIN X X DOUBLE X X DR X DROP X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 182 SQL Reference: Fundamentals DUAL X DUMP X DYNAMIC X DYNAMIC_FUNCTION X DYNAMIC_FUNCTION_CODE X EACH X X EBCDIC X ECHO X ELEMENT X ELSE X X ELSEIF X ENABLED X ENCRYPT X END X X END-EXEC X EQ X EQUALS X X ERROR X ERRORFILES X ERRORTABLES X ESCAPE X X ET X EVERY X EXCEPT X X EXCEPTION X EXCL X EXCLUDE X EXCLUDING X EXCLUSIVE X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 183 EXEC X X EXECUTE X X EXISTING EXISTS X X EXIT X EXP X X EXPIRE X EXPLAIN X EXTERNAL X X EXTRACT X X FALLBACK X FALSE X FASTEXPORT X FETCH X X FILTER X FINAL X X FIRST X X FLOAT X X FLOOR X FOLLOWING X X FOR X X FOREIGN X X FORMAT X FORTRAN X FOUND X X FREE X FREESPACE X FROM X X FULL X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 184 SQL Reference: Fundamentals FUNCTION X X FUSION X G X X GE X GENERAL X GENERATED X X GET X GIVE X GLOBAL X X GO X X GOTO X X GRANT X X GRANTED X GRAPHIC X GROUP X X GROUPING X X GT X HANDLER X HASH X HASHAMP X HASHBAKAMP X HASHBUCKET X HASHROW X HAVING X X HELP X HIERARCHY X HIGH X HOLD X HOST X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 185 HOUR X X IDENTITY X X IF X IFP X IMMEDIATE X X IMPLEMENTATION X IN X X INCLUDING X INCONSISTENT X INCREMENT X X INDEX X INDEXESPERTABLE X INDEXMAINTMODE X INDICATOR X X INITIALLY X INITIATE X INNER X X INOUT X X INPUT X X INS X INSENSITIVE X INSERT X X INSTANCE X X INSTANTIABLE X X INSTEAD X INT X X INTEGER X X INTEGERDATE X INTERSECT X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 186 SQL Reference: Fundamentals INTERSECTION X INTERVAL X X INTO X X INVOKER X X IOCOUNT X IS X X ISOLATION X X ITERATE X JAVA X JIS_COLL X JOIN X X JOURNAL X K X X KANJI1 X KANJISJIS X KBYTE X KBYTES X KEEP X KEY X X KEY_MEMBER X KEY_TYPE X KILOBYTES X KURTOSIS X LANGUAGE X X LARGE X X LAST X X LATERAL X LATIN X LE X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 187 Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReserved LEADING X X LEAVE X LEFT X X LENGTH X LEVEL X X LIKE X X LIMIT X LN X X LOADING X LOCAL X X LOCALTIME X LOCALTIMESTAMP X LOCATOR X X LOCK X LOCKEDUSEREXPIRE X LOCKING X LOG X LOGGING X LOGON X LONG X LOOP X LOW X LOWER X X LT X M X X MACRO X MAP X X MATCH XAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 188 SQL Reference: Fundamentals MATCHED X X MAVG X MAX X X MAXCHAR X MAXIMUM X MAXLOGONATTEMPTS X MAXVALUE X X MCHARACTERS X MDIFF X MEDIUM X MEMBER X MERGE X X MESSAGE_LENGTH X MESSAGE_OCTET_LENGTH X MESSAGE_TEXT X METHOD X X MIN X X MINCHAR X MINDEX X MINIMUM X MINUS X MINUTE X X MINVALUE X X MLINREG X MLOAD X MOD X X MODE X MODIFIED X MODIFIES X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 189 MODIFY X MODULE X MONITOR X MONRESOURCE X MONSESSION X MONTH X X MORE X MSUBSTR X MSUM X MULTINATIONAL X MULTISET X X MUMPS X NAME X X NAMED X NAMES X NATIONAL X NATURAL X X NCHAR X NCLOB X NE X NESTING X NEW X X NEW_TABLE X NEXT X X NO X X NONE X X NORMALIZE X NORMALIZED X NOT X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 190 SQL Reference: Fundamentals NOWAIT X NULL X X NULLABLE X NULLIF X X NULLIFZERO X NULLS X NUMBER X NUMERIC X X OA X OBJECT X X OBJECTS X OCTET_LENGTH X X OCTETS X OF X X OFF X OLD X X OLD_TABLE X ON X X ONLY X X OPEN X X OPTION X X OPTIONS X OR X X ORDER X X ORDERED_ANALYTIC X ORDERING X X ORDINALITY X OTHERS X OUT X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 191 OUTER X X OUTPUT X OVER X X OVERLAPS X X OVERLAY X OVERRIDE X OVERRIDING X PAD X PARAMETER X X PARAMETER_MODE X PARAMETER_NAME X PARAMETER_ORDINAL_POSITION X PARAMETER_SPECIFIC_CATALOG X PARAMETER_SPECIFIC_NAME X PARAMETER_SPECIFIC_SCHEMA X PARTIAL X PARTITION X X PARTITIONED X PASCAL X PASSWORD X PATH X PERCENT X PERCENT_RANK X X PERCENTILE_CONT X PERCENTILE_DISC X PERM X PERMANENT X PLACING X PLI X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 192 SQL Reference: Fundamentals POSITION X X POWER X PRECEDING X X PRECISION X X PREPARE X X PRESERVE X X PRIMARY X X PRINT X PRIOR X PRIVATE X PRIVILEGES X X PROCEDURE X X PROFILE X PROTECTED X PROTECTION X PUBLIC X X QUALIFIED X QUALIFY X QUANTILE X QUEUE X QUERY X RADIANS X RANDOM X RANDOMIZED X RANGE X X RANGE_N X RANK X X READ X X READS X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 193 REAL X X RECALC X RECURSIVE X X REF X REFERENCES X X REFERENCING X X REGR_AVGX X X REGR_AVGY X X REGR_COUNT X X REGR_INTERCEPT X X REGR_R2 X X REGR_SLOPE X X REGR_SXX X X REGR_SXY X X REGR_SYY X X RELATIVE X X RELEASE X X RENAME X REPEAT X REPEATABLE X REPLACE X REPLACEMENT X REPLCONTROL X REPLICATION X REQUEST X RESTART X X RESTORE X RESTRICT X RESULT X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 194 SQL Reference: Fundamentals RESUME X RET X RETAIN X RETRIEVE X RETURN X RETURNED_CARDINALITY X RETURNED_LENGTH X RETURNED_OCTET_LENGTH X RETURNED_SQLSTATE X RETURNS X X REUSE X REVALIDATE X REVOKE X X RIGHT X X RIGHTS X ROLE X X ROLLBACK X X ROLLFORWARD X ROLLUP X X ROUTINE X ROUTINE_CATALOG X ROUTINE_NAME X ROUTINE_SCHEMA X ROW X X ROW_COUNT X ROW_NUMBER X X ROWID X ROWS X X RU X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 195 SAMPLE X SAMPLEID X SAMPLES X SAVEPOINT X SCALE X SCHEMA X SCHEMA_NAME X SCOPE X SCOPE_CATALOG X SCOPE_NAME X SCOPE_SCHEMA X SCROLL X X SEARCH X SEARCHSPACE X SECOND X X SECTION X SECURITY X X SEED X SEL X SELECT X X SELF X X SENSITIVE X SEQUENCE X SERIALIZABLE X X SERVER_NAME X SESSION X X SESSION_USER X SET X X SETRESRATE X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 196 SQL Reference: Fundamentals SETS X X SETSESSRATE X SHARE X SHOW X SIMILAR X SIMPLE X SIN X SINH X SIZE X SKEW X SMALLINT X X SOME X X SOUNDEX X SOURCE X X SPACE X SPECCHAR X SPECIFIC X X SPECIFIC_NAME X SPECIFICTYPE X SPL X SPOOL X SQL X X SQLEXCEPTION X X SQLSTATE X X SQLTEXT X SQLWARNING X X SQRT X X SR X SS X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 197 START X X STARTUP X STAT X STATE X STATEMENT X X STATIC X STATISTICS X STATS X STDDEV_POP X X STDDEV_SAMP X X STEPINFO X STRING_CS X STRUCTURE X STYLE X X SUBCLASS_ORIGIN X SUBLIST SUBMULTISET X SUBSCRIBER X SUBSTR X SUBSTRING X X SUM X X SUMMARY X SUMMARYONLY X SUSPEND X SYMMETRIC X SYSTEM X X SYSTEM_USER X SYSTEMTEST X TABLE X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 198 SQL Reference: Fundamentals TABLE_NAME X TABLESAMPLE X TAN X TANH X TARGET X TBL_CS X TD_GENERAL X TD_INTERNAL X TEMPORARY X X TERMINATE X TEXT X THAN THEN X X THRESHOLD X TIES X X TIME X X TIMESTAMP X X TIMEZONE_HOUR X X TIMEZONE_MINUTE X X TITLE X TO X X TOP X TPA X TOP_LEVEL_COUNT X TRACE X TRAILING X X TRANSACTION X X TRANSACTION_ACTIVE X TRANSACTIONS_COMMITTED X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 199 TRANSACTIONS_ROLLED_BACK X TRANSFORM X X TRANSFORMS X TRANSLATE X X TRANSLATE_CHK X TRANSLATION X TREAT X TRIGGER X X TRIGGER_CATALOG X TRIGGER_NAME X TRIGGER_SCHEMA X TRIM X X TRUE X TYPE X X UC X UDTCASTAS X UDTCASTLPAREN X UDTMETHOD X UDTTYPE X UDTUSAGE X UESCAPE X UNBOUNDED X X UNCOMMITTED X X UNDEFINED X UNDER X UNDO X UNICODE X UNION X X UNIQUE X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 200 SQL Reference: Fundamentals UNKNOWN X X UNNAMED X UNNEST X UNTIL X UPD X UPDATE X X UPPER X X UPPERCASE X USAGE X USE X USER X X USER_DEFINED_TYPE_CATALOG X USER_DEFINED_TYPE_CODE X USER_DEFINED_TYPE_NAME X USER_DEFINED_TYPE_SCHEMA X USING X X VALUE X X VALUES X X VAR_POP X X VAR_SAMP X X VARBYTE X VARCHAR X X VARGRAPHIC X VARYING X X VIEW X X VOLATILE X WAIT X WARNING X WHEN X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 SQL Reference: Fundamentals 201 WHENEVER X WHERE X X WHILE X WIDTH_BUCKET X X WINDOW X WITH X X WITHIN X WITHOUT X WORK X X WRITE X X YEAR X X ZEROIFNULL X ZONE X X Keyword Teradata Database Status SQL-2003 Status Reserved Future NonReserved Reserved NonReservedAppendix B: Restricted Words for V2R6.2 Reserved Words and Keywords for V2R6.2 202 SQL Reference: FundamentalsSQL Reference: Fundamentals 203 APPENDIX C Teradata Database Limits This appendix provides the following Teradata Database limits: • System limits • Database limits • Session limitsAppendix C: Teradata Database Limits System Limits 204 SQL Reference: Fundamentals System Limits The system specifications in the following table apply to an entire Teradata Database configuration. Parameter Value Maximum number of databases and users 4.2 x 10 9 Total data capacity • Expressed as a base 10 value: • 1.39 TB/AMP (1.39 x 10 12 bytes/AMP) • Expressed as a base 2 value: • 1.26 TB/AMP (1.26 x 10 12 bytes/AMP) Maximum number of active concurrent transactions 2048 Maximum data format descriptor size 30 characters Maximum error message text size in failure parcel 255 bytes Maximum number of sectors per datablock 255 a Maximum data block size 130560 bytes Datablock header size Depends on several factors: FOR a datablock that is … The datablock header size is this many bytes … new or has been updated 72 on a 64-bit system and has not been updated 40 on a 32-bit system and has not been updated 36 Maximum number of sessions per PE 120 Maximum number of gateways per node 1 Maximum number of sessions per Gateway Tunable. b 1200 maximum certified Maximum number of parcels in one message 256Appendix C: Teradata Database Limits System Limits SQL Reference: Fundamentals 205 Maximum message size Approximately 65000 bytes Note: This limit applies to messages to/from host systems and to some internal Teradata Database messages. Maximum number of PEs per system 1024 Maximum number of AMPs per system 16383 c More generally, the maximum number of AMPs per system depends on the number of PEs in the configuration. The following equation provides the most general solution: Maximum number of AMP and PE vprocs, in any combination, per system 16384 Number of hash buckets per system 65536 d Bucket numbers range from 0 to 65535. Number of hash values per system 4.2 x 10 9 Maximum number of external routine protected mode server tasks per PE or AMP 20 e Maximum number of external routine secure mode server tasks per PE or AMP 20 c Amount of private disk swap space required per protected or secure mode server per PE or AMP vproc 256 KB a. The increase in datablock header size from 36 or 40 bytes to 64 bytes increases the size of roughly 6 percent of the datablocks by one sector (see “Datablock header size” on page 204). b. See Utilities for details. c. This value is derived by subtracting 1 from the maximum total of PE and AMP vprocs per system (because each system must have at least one PE), which is 16384. This is obviously not a practical configuration. d. This value is fixed. The system assigns its 65536 hash buckets to AMPs as evenly as possible. For example, a system with 1000 AMPs has 65 hash buckets on some AMPs and 66 hash buckets on others. In this particular case, the AMPs having 66 hash buckets also perform 1.5 percent more work than those with 65 hash buckets. The work per AMP imbalance increases as a function of the number of AMPs in the system for those cases where 65536 is not evenly divisible by the total number of AMPs. e. The valid range is 0 to 20, inclusive. The limit is 20 servers for each server type, not 20 combined for both. See Utilities for details. Parameter Value 16384 number_of_PEs –Appendix C: Teradata Database Limits Database Limits 206 SQL Reference: Fundamentals Database Limits The database specifications in the following table apply to a single database. The values presented are maxima for their respective parameters individually and not in combination. Parameter Maximum Value Number of journal tables per database 1 Number of data tables per database 4.2 x 10 9 Database, user, table, view, macro, index, constraint, userdefined function, stored procedure, user-defined method, user-defined type, replication group, or column name size 30 bytes Tables and Views Number of columns per base table or view 2048 Number of UDT columns per base table or view Approximately 1600 g,h,i Number of LOB type columns per base table 32 j Number of columns created over the life of a base table 2560 Number of rows per base table Limited by disk capacity Number of bytes per table header a • Approximately 64000 bytes -or- • Approximately 128000 bytes Row size Approximately 65536 bytes Logical row size b 67106816000 bytes k Number of secondary c , hash, and join indexes, in any combination, per table 32 Non-LOB column size • 65522 bytes (NPPI table) l • 65520 bytes (PPI table)m Number of columns per primary or secondary index 64 SQL title size 60 characters Size of the queue table FIFO runtime cache per PE • 100 queue table entries • 1 MB Size of the queue table FIFO runtime cache per table 2211 row entries Number of primary indexes per table 1 Number of partitions for a partitioned primary index 65535 Number of table-level constraints per table 100Appendix C: Teradata Database Limits Database Limits SQL Reference: Fundamentals 207 Number of referential constraints per table 64 Number of columns in foreign and parent keys 64 Number of compressed values per column 255 plus nulls Predefined and User-Defined Types BLOB object size 2097088000 bytes CLOB object size • 2097088000 single-byte characters • 1048544000 double-byte characters Structured UDT size. d • 65521 bytes (NPPI table) • 65519 bytes (PPI table) Number of characters in a string constant 32000 Number of attributes that can be specified for a structured UDT per CREATE TYPE or ALTER TYPE statement 300 - 512 n Number of attributes that can be defined for a structured UDT Approximately 4000 o Number of nested attributes in a structured UDT 512 Number of methods associated with a UDT Approximately 500 p Macros, Stored Procedures, and External Routines Expanded text size for macros and views 2 MB Length of external name string for an external routine. e 1000 characters Package path length for an external routine 256 characters SQL request size in a stored procedure 64 KB Number of parameters specified in a UDF 128 Number of parameters specified in a UDM 128 Number of parameters specified in a macro 2048 Number of parameters in a stored procedure 256 Number of nested CALL statements 15 Number of open cursors 16 for embedded SQL, 15 for a stored procedure Queries, Requests, and Responses SQL request size 1 MB (Includes SQL statement text, USING data, and parcel overhead) Parameter Maximum ValueAppendix C: Teradata Database Limits Database Limits 208 SQL Reference: Fundamentals SQL response size 1 MB (Includes SQL result and parcel overhead) Number of columns per DML statement ORDER BY clause 16 Number of tables that can be joined per query block 64 Number of subquery nesting levels per query 64 Number of fields in a USING row descriptor 2550 SQL activity count size 2 32 -1 rows Number of SELECT AND CONSUME statements in a delayed state per PE 24 Number of partitions for a hash join operation 50 Query and Workload Analysis Size of the Index Wizard workload cache 256 MB q Number of indexes on which statistics can be collected and maintained at one time 32 This limit is independent of the number of pseudoindexes on which statistics can be collected and maintained. Number of pseudoindexes f on which multicolumn statistics can be collected and maintained at one time 32 This limit is independent of the number of indexes on which statistics can be collected and maintained. Number of columns and indexes on which statistics can be recollected for a table 512 Hash and Join Indexes Number of columns referenced per single table in a hash or join index 64 Number of columns referenced in the fixed part of a compressed hash or join index 64 Number of columns referenced in the repeating part of a compressed hash or join index 64 Number of columns in an uncompressed join index 2048 Number of columns in a compressed join index 128 Parameter Maximum ValueAppendix C: Teradata Database Limits Database Limits SQL Reference: Fundamentals 209 Replication Row size permitted for a replication operation Approximately 25000 bytes For details, see Teradata Replication Solutions Overview and “CREATE REPLICATION GROUP” in SQL Reference: Data Definition Statements. Number of replication groups per system 100 Number of tables that can be copied simultaneously with a replication operation 15 Number of columns that can be defined for a replicated table 1000 Character column data size permitted for a replication operation • CHARACTER(10000) • VARCHAR(10000) For UTF16, this translates to a maximum of 5000 characters. a. A table header that is large enough to require more than ~64000 bytes uses two 64Kbyte rows. A table header that requires 64000 or fewer bytes does not use the second row that is required to contain a table header of ~128000 bytes. b. A logical row is defined as a base table row plus the sum of the bytes stored in a LOB subtable for that row. c. A NUSI defined with an ORDER BY clause counts as two indexes in this calculation. d. Based on a table having a 1 byte (BYTEINT) primary index. Because a UDT column cannot be part of any index definition, there must be at least one non-UDT column in the table for its primary index. Row header overhead consumes 14 bytes in an NPPI table and 16 bytes in a PPI table, so the maximum structured UDT size is derived by subtracting 15 bytes (for an NPPI table) or 17 bytes (for a PPI table) from the row maximum of 65 536 bytes. e. An external routine is the portion of a UDF, external stored procedure, or method that is written in C or C++. This is the code that defines the semantics for the UDF, procedure, or method. f. A pseudoindex is a file structure that allows you to collect statistics on a composite, or multicolumn, column set in the same way you collect statistics on a composite index. g. The absolute limit is 2048, and the realizable number varies as a function of the number of other features declared for a table that occupy table header space. h. The figure of 1600 UDT columns assumes a FAT table header. i. This limit is true whether the UDT is a distinct or a structured type. j. This includes both predefined type LOB columns and UDT LOB columns. A UDT LOB column counts as one LOB column even if the UDT is a structured type that has multiple LOB attributes. k. This value is derived by multiplying the maximum number of LOB columns per base table (32) times the maximum size of a LOB field (2 097 088 000 8-bit bytes). Remember that each LOB column consumes 39 bytes of Object ID from the base table, so 1 248 of those 67 106 816 000 bytes cannot be used for data. l. Based on subtracting the minimum row overhead value for an NPPI table row (14 bytes) from the system-defined maximum row length (65 536 bytes). Parameter Maximum ValueAppendix C: Teradata Database Limits Database Limits 210 SQL Reference: Fundamentals m. Based on subtracting the minimum row overhead value for a PPI table row (16 bytes) from the system-defined maximum row length (65 536 bytes). n. The maximum is platform-dependent. o. While you can specify no more than 300 to 512 attributes for a structured UDT per CREATE TYPE or ALTER TYPE statement, you can submit any number of ALTER TYPE statements with the ADD ATTRIBUTE option specified as necessary to add additional attributes to the type up to the upper limit of approximately 4000. p. There is no absolute limit on the number of methods that can be associated with a given UDT. Methods can have a variable number of parameters, and the number of parameters directly affects the limit, which is due to parser memory restrictions. There is a workaround for this issue. See the documentation for ALTER TYPE in SQL Reference: Data Manipulation Statements for details. q. The default is 48 megabytes and the minimum is 32 megabytes.Appendix C: Teradata Database Limits Session Limits SQL Reference: Fundamentals 211 Session Limits The session specifications in the following table apply to a single session. Parameter Value Active request result spool files 16 Parallel steps Parallel steps can be used to process a request submitted within a transaction (which may be either explicit or implicit). The maximum number of steps generated per request is determined as follows: • Per request, if no channels Note: Channels are not required for a primary index request with an equality constraint. 20 steps • A request that involves redistribution of rows to other AMPs, such as a join or an INSERT-SELECT Requires 4 channels • A request that does not involve row distribution Requires 2 channels Number of materialized global temporary tables per session 2000 Number of volatile tables per session 1000Appendix C: Teradata Database Limits Session Limits 212 SQL Reference: FundamentalsSQL Reference: Fundamentals 213 APPENDIX D ANSI SQL Compliance This appendix describes the ANSI SQL standard, Teradata compliance with the ANSI SQL standard, and terminology differences between ANSI SQL and Teradata SQL. Topics include: • ANSI SQL Standard • Terminology Differences Between ANSI SQL and Teradata • SQL Flagger • Differences Between Teradata and ANSI SQL ANSI SQL Standard Introduction The American National Standards Institute (ANSI) SQL standard, formally titled International Standard ISO/IEC 9075:2003, Database Language SQL, defines a version of Structured Query Language that all vendors of relational database management systems support to a greater or lesser degree. Motivation Behind an SQL Standard Teradata, like most vendors of relational database management systems, had its own dialect of the SQL language for many years prior to the development of the SQL standard. You might ask several questions like the following: • Why should there be an industry-wide SQL standard? • Why should any vendor with an entrenched user base consider modifying its SQL dialect to conform with the ANSI SQL standard? Why an SQL Standard? National and international standards abound in the computer industry. As anyone who has worked in the industry for any length of time knows, standardization offers both advantages and disadvantages both to users and to vendors. The principal advantages of having an SQL standard are the following: • Open systems The overwhelming trend in computer technology has been toward open systems with publicly defined standards to facilitate third party and end user access and development using the standardized products.Appendix D: ANSI SQL Compliance ANSI SQL Standard 214 SQL Reference: Fundamentals The ANSI SQL standard provides an open definition for the SQL language. • Less training for transfer and new employees A programmer trained in ANSI-standard SQL can move from one SQL programming environment to another with no need to learn a new SQL dialect. When a core dialect of the language is the lingua franca for SQL programmers, the need for retraining is significantly reduced. • Application portability When there is a standardized public definition for a programming language, users can rest assured that any applications they develop to the specifications of that standard are portable to any environment that supports the same standard. This is an extremely important budgetary consideration for any large scale end user application development project. • Definition and manipulation of heterogeneous databases is facilitated Many user data centers support multiple merchant databases across different platforms. A standard language for communicating with relational databases, irrespective of the vendor offering the database management software, is an important factor in reducing the overhead of maintaining such an environment. • Intersystem communication is facilitated It is common for an enterprise to exchange applications and data among different merchant databases. Common examples of this appear below. • Two-phase commit transactions where rows are written to multiple databases simultaneously. • Bulk data import and export between different vendor databases. These operations are made much cleaner and simpler when there is no need to translate data types, database definitions, and other component definitions between source and target databases. Teradata Compliance With the ANSI Standard Conformance to a standard presents problems for any vendor that produces an evolved product and supports a large user base. Teradata, in its historical development, has produced any number of innovative SQL language elements that do not conform to the ANSI SQL standard, a standard that did not exist when those features were conceived. The existing Teradata user base had invested substantial time, effort, and capital into developing applications using that Teradata SQL dialect. At the same time, new customers demand that vendors conform to open standards for everything from chip sets to operating systems to application programming interfaces. Meeting these divergent requirements presents a challenge that Teradata SQL solves by following the multipronged policy outlined in the following table.Appendix D: ANSI SQL Compliance ANSI SQL Standard SQL Reference: Fundamentals 215 WHEN … THEN … a new feature or feature enhancement is added to Teradata SQL that feature conforms to the ANSI SQL standard. the difference between the Teradata SQL dialect and the ANSI SQL standard for a language feature is slight the ANSI SQL is added to the Teradata Database feature as an option. the difference between the Teradata SQL dialect and the ANSI SQL standard for a language feature is significant both syntaxes are offered and the user has the choice of operating in either Teradata or ANSI mode or of turning off SQL Flagger. The mode can be defined in the following ways: • Persistently Use the SessionMode field of the DBS Control Record to define session mode characteristics. • For a session Use the BTEQ .SET SESSION TRANSACTION command to control transaction semantics. Use the BTEQ .SET SESSION SQLFLAG command to control use of the SQL Flagger. Use the SQL statement SET SESSION DATEFORM to control how data typed as DATE is handled. a new feature or feature enhancement is added to Teradata SQL and that feature is not defined by the ANSI SQL standard that feature is designed using the following criteria: IF other vendors … THEN Teradata designs the new feature … offer a similar feature or feature extension to broadly comply with other solutions, but consolidates the best ideas from all and, where necessary, creates its own, cleaner solution. do not offer a similar feature or feature extension • as cleanly and generically as possible with an eye toward creating a language element that will not be subject to major revisions to comply with future updates to the ANSI SQL standard. • in a way that offers the most power to users without violating any of the basic tenets of the ANSI SQL standard.Appendix D: ANSI SQL Compliance Terminology Differences Between ANSI SQL and Teradata 216 SQL Reference: Fundamentals Terminology Differences Between ANSI SQL and Teradata The ANSI SQL standard and Teradata occasionally use different terminology. The following table lists the more important variances. Note: 1) In the ANSI SQL standard, the term table has the following definitions: • A base table • A viewed table (view) • A derived table ANSI Teradata Base table Table 1 Binding style Not defined, but implicitly includes the following: • Interactive SQL • Embedded SQL • ODBC • CLIv2 Authorization ID User ID Catalog Dictionary CLI ODBC 2 Direct SQL Interactive SQL Domain Not defined External routine function User-defined function (UDF) Module Not defined Persistent stored module Stored procedure Schema User Database SQL database Relational database Viewed table View Not defined Explicit transaction 3 Not defined CLIv2 4 Not defined Macro 5Appendix D: ANSI SQL Compliance SQL Flagger SQL Reference: Fundamentals 217 2) ANSI CLI is not exactly equivalent to ODBC, but the ANSI standard is heavily based on the ODBC definition. 3) ANSI transactions are always implicit, beginning with an executable SQL statement and ending with either a COMMIT or a ROLLBACK statement. 4) Teradata CLIv2 is an implementation-defined binding style. 5) The function of Teradata Database macros is similar to that of ANSI persistent stored modules without having the loop and branch capabilities stored modules offer. SQL Flagger Function The SQL Flagger, when enabled, reports the use of non-standard SQL. The SQL Flagger always permits statements flagged as non-entry-level or noncompliant ANSI SQL to execute. Its task is not to enforce the standard, but rather to return a warning message to the requestor noting the noncompliance. The analysis includes syntax checking as well as some dictionary lookup, particularly the implicit assignment and comparison of different data types (where ANSI requires use of the CAST function to convert the types explicitly) as well as some semantic checks. The SQL Flagger does not check or detect every condition for noncompliance; thus, a statement that is not flagged does not necessarily mean it is compliant. Enabling and Disabling the SQL Flagger Flagging is enabled by a client application before a session is logged on and generally is used only to assist in checking for ANSI compliance in code that must be portable across multiple vendor environments. The SQL Flagger is disabled by default. You can enable or disable it using any of the following procedures, depending on your application. FOR this software … USE these commands or options … TO turn the SQL Flagger … BTEQ .[SET] SESSION SQLFLAG ENTRY to entry-level ANSI .[SET] SESSION SQLFLAG NONE off See Basic Teradata Query Reference for more detail on using BTEQ commands. Preprocessor2 SQLFLAGGER(ENTRY) to entry-level ANSI SQLFLAGGER(NONE) off See Teradata Preprocessor2 for Embedded SQL Programmer Guide for details on setting Preprocessor options.Appendix D: ANSI SQL Compliance Differences Between Teradata and ANSI SQL 218 SQL Reference: Fundamentals Differences Between Teradata and ANSI SQL For a complete list of SQL features in this release, see Appendix E. The list identifies which features are ANSI SQL compliant and which features are Teradata extensions. The list of features includes SQL statements and options, functions and operators, data types and literals. CLI set lang_conformance = ‘2’ set lang_conformance to ‘2’ to entry-level ANSI set lang_conformance = ‘N’ off See Teradata Call-Level Interface Version 2 Reference for Channel-Attached Systems and Teradata Call-Level Interface Version 2 Reference for NetworkAttached Systems for details on setting the conformance field. FOR this software … USE these commands or options … TO turn the SQL Flagger …SQL Reference: Fundamentals 219 APPENDIX E SQL Feature Summary This appendix details the differences in SQL between this release and previous releases. • “Statements and Modifiers” on page 219 • “Data Types and Literals” on page 277 • “Functions, Operators, and Expressions” on page 280 The intent of this appendix is to provide a way to readily identify new SQL in this release and previous releases of Teradata Database. It is not meant as a Teradata SQL reference. Notation Conventions The following table describes the conventions used in this appendix. Statements and Modifiers The following table lists SQL statements and modifiers for this version and previous versions of Teradata Database. The following type codes appear in the ANSI Compliance column. This notation … Means … UPPERCASE a keyword italics a variable, such as a column or table name [ n ] that the use of n is optional | n | that option n is described separately in this appendix Code Definition A ANSI SQL-2003 compliant T Teradata extensionAppendix E: SQL Feature Summary Statements and Modifiers 220 SQL Reference: Fundamentals Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0 ABORT T X X X Options FROM option T X X X WHERE condition T X X X ALTER FUNCTION ALTER SPECIFIC FUNCTION T X X X Options EXECUTE PROTECTED/ EXECUTE NOT PROTECTED T X X X COMPILE/ COMPILE ONLY T X X X ALTER METHOD ALTER CONSTRUCTOR METHOD ALTER INSTANCE METHOD ALTER SPECIFIC METHOD T X X Options EXECUTE PROTECTED/ EXECUTE NOT PROTECTED/ COMPILE/ COMPILE ONLY T X X ALTER PROCEDURE (external form) T X X X Options LANGUAGE C/ LANGUAGE CPP T X X X COMPILE/ COMPILE ONLY/ EXECUTE PROTECTED/ EXECUTE NOT PROTECTED T X X XAppendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 221 ALTER PROCEDURE (internal form) T X X X Options COMPILE T X X X WITH PRINT/ WITH NO PRINT T X X X WITH SPL/ WITH NO SPL T X X X WITH WARNING/ WITH NO WARNING T X X X ALTER REPLICATION GROUP T X X X Options ADD table_name/ ADD database_name.table_name T X X X DROP table_name/ DROP database_name.table_name T X X X ALTER TABLE A, T X X X Options ADD column_name |Data Type| |Data Type Attributes| A X X X ADD column_name |Column Storage Attributes| T X X X ADD column_name NO COMPRESS T X ADD column_name |Column Constraint Attributes| T X X X ADD column_name |Table Constraint Attributes| T X X X ADD |Table Constraint Attributes| A X X X ADD column_name NULL T X X X AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL AFTER JOURNAL/ NOT LOCAL AFTER JOURNAL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 222 SQL Reference: Fundamentals ALTER TABLE, continued Options BEFORE JOURNAL/ JOURNAL/ NO BEFORE JOURNAL/ DUAL BEFORE JOURNAL T X X X DATABLOCKSIZE IMMEDIATE/ MINIMUM DATABLOCKSIZE/ MAXIMUM DATABLOCKSIZE/ DEFAULT DATABLOCKSIZE T X X X CHECKSUM = DEFAULT/ CHECKSUM = NONE/ CHECKSUM = LOW/ CHECKSUM = MEDIUM/ CHECKSUM = HIGH/ CHECKSUM = ALL T X X X DROP column_name A X X X DROP CHECK/ DROP column_name CHECK/ DROP CONSTRAINT name CHECK T X X X DROP CONSTRAINT T X X X DROP FOREIGN KEY REFERENCES T X X X WITH CHECK OPTION/ WITH NO CHECK OPTION T X X X DROP INCONSISTENT REFERENCES T X X X FALLBACK PROTECTION/ NO FALLBACK PROTECTION T X X X FREESPACE/ DEFAULT FREESPACE T X X X LOG/NO LOG T X X X MODIFY CHECK/ MODIFY column_name CHECK/ MODIFY CONSTRAINT name CHECK T X X X MODIFY [[NOT] UNIQUE] PRIMARY INDEX index [(column)]/ MODIFY [[NOT] UNIQUE] PRIMARY INDEX NOT NAMED [(column)] T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 223 ALTER TABLE, continued Options NOT PARTITIONED/ PARTITION BY expression/ DROP RANGE WHERE expression [ADD RANGE ranges]/ DROP RANGE ranges [ADD RANGE ranges]/ ADD RANGE ranges T X X X ON COMMIT DELETE ROWS/ ON COMMIT PRESERVE ROWS T X X X RENAME column_name T X X X REVALIDATE PRIMARY INDEX/ REVALIDATE PRIMARY INDEX WITH DELETE/ REVALIDATE PRIMARY INDEX WITH INSERT [INTO] table_name T X X X WITH JOURNAL TABLE T X X X ALTER TRIGGER T X X X Options ENABLED/DISABLED / T X X X TIMESTAMP T X X X ALTER TYPE A,T X X Options ADD ATTRIBUTE/ DROP ATTRIBUTE/ ADD METHOD/ ADD INSTANCE METHOD/ ADD CONSTRUCTOR METHOD/ ADD SPECIFIC METHOD/ DROP METHOD/ DROP INSTANCE METHOD/ DROP CONSTRUCTOR METHOD/ DROP SPECIFIC METHOD A,T X X BEGIN DECLARE SECTION A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 224 SQL Reference: Fundamentals BEGIN LOGGING T X X X Options DENIALS T X X X WITH TEXT T X X X FIRST/ LAST/ FIRST AND LAST/ EACH T X X X BY database_name T X X X ON ALL/ ON operation/ ON GRANT T X X X ON DATABASE/ ON USER/ ON TABLE/ ON VIEW/ ON MACRO/ ON PROCEDURE/ ON FUNCTION/ T X X X ON TYPE T X X BEGIN QUERY LOGGING T X X X Options WITH ALL/ WITH OBJECTS/ WITH SQL/ WITH STEPINFO/ T X X X WITH COSTS T X X X LIMIT SQLTEXT [=n] [AND …]/ LIMIT SUMMARY = n1, n2, n3 [AND …]/ LIMIT THRESHOLD [=n] [AND …]/ T X X X LIMIT MAXCPU [=n] [AND …] T X X X ON ALL/ ON user_name/ ON user_name ACCOUNT = 'account_name'/ ON user_name ACCOUNT = ('account_name' [ … ,'account_name']) T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 225 BEGIN TRANSACTION/ BT T X X X CALL A X X X Options stored_procedure_name/ A X X X external_stored_procedure_name A X X X CHECKPOINT T X X X Options NAMED checkpoint T X X X INTO host_variable_name T X X X [INDICATOR] :host_indicator_name T X X X CLOSE A X X X COLLECT DEMOGRAPHICS T X X X Options FOR table_name/ FOR (table_name [ … ,table_name]) T X X X ALL/ WITH NO INDEX T X X X COLLECT STATISTICS/ COLLECT STATS/ COLLECT STAT (QCD form) T X X X Options PERCENT T X X X SET QUERY query_ID T X X SAMPLEID statistics_ID T X X UPDATE MODIFIED T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 226 SQL Reference: Fundamentals COLLECT STATISTICS (QCD form), continued Options INDEX (column_name [ … , column_name])/ INDEX index_name/ COLUMN (column_name [ … ,column_name])/ COLUMN column_name/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X COLLECT STATISTICS/ COLLECT STATS/ COLLECT STAT (optimizer form) T X X X Options USING SAMPLE T X X X [ON] [TEMPORARY] table_name/ [ON] join_index_name/ [ON] hash_index_name T X X X INDEX (column_name [ … , column_name])/ INDEX index_name/ COLUMN (column_name [ … ,column_name])/ COLUMN column_name/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 227 COLLECT STATISTICS/ COLLECT STATS/ COLLECT STAT (optimizer form, CREATE INDEX-style syntax) T X X X Options USING SAMPLE T X X X [UNIQUE] INDEX [index_name] [ALL] (column_name [ … , column_name]) [ORDER BY [VALUES]] (column_name)/ [UNIQUE] INDEX [index_name] [ALL] (column_name [ … , column_name]) [ORDER BY [HASH]] (column_name)/ COLUMN column_name/ COLUMN (column_name [ … , column_name])/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X ON [TEMPORARY] table_name/ ON hash_index_name/ ON join_index_name T X X X COMMENT T X X X Options [ON] COLUMN object_name/ [ON] DATABASE object_name/ [ON] FUNCTION object_name/ [ON] MACRO object_name/ [ON] PROCEDURE object_name/ [ON] TABLE object_name/ [ON] TRIGGER object_name/ [ON] USER object_name/ [ON] VIEW object_name/ [ON] PROFILE object_name/ [ON] ROLE object_name/ T X X X [ON] GROUP group_name/ T X X X [ON] METHOD object_name/ [ON] TYPE object_name T X X AS 'comment'/ IS 'comment' T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 228 SQL Reference: Fundamentals COMMENT (embedded SQL) T X X X Options [ON] COLUMN object_reference/ [ON] DATABASE object_reference/ [ON] FUNCTION object_name/ [ON] MACRO object_reference/ [ON] PROCEDURE object_reference/ [ON] TABLE object_reference/ [ON] TRIGGER object_reference/ [ON] USER object_reference/ [ON] VIEW object_reference/ [ON] PROFILE object_name/ [ON] ROLE object_name/ T X X X [ON] GROUP group_name T X X X INTO host_variable_name T X X X [INDICATOR] :host_indicator_name T X X X COMMIT A, T X X X Options WORK A X X X RELEASE T X X X CONNECT (embedded SQL) T X X X Options IDENTIFIED BY passwordvar/ IDENTIFIED BY :passwordvar T X X X AS connection_name/ AS :namevar T X X X CREATE AUTHORIZATION T X X Options [AS] DEFINER/ [AS] DEFINER DEFAULT/ [AS] INVOKER T X X DOMAIN 'domain_name' T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 229 CREATE CAST A X X Options WITH SPECIFIC METHOD specific_method_name/ WITH METHOD method_name/ WITH INSTANCE METHOD method_name/ WITH SPECIFIC FUNCTION specific_function_name/ WITH FUNCTION function_name A X X AS ASSIGNMENT A X X CREATE DATABASE T X X X Options PERMANENT = n [BYTES] T X X X SPOOL = n [BYTES] T X X X TEMPORARY = n [BYTES] T X X X ACCOUNT T X X X FALLBACK [PROTECTION]/ NO FALLBACK [PROTECTION] T X X X BEFORE JOURNAL/ JOURNAL/ NO JOURNAL NO BEFORE JOURNAL/ DUAL JOURNAL DUAL BEFORE JOURNAL T X X X AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL AFTER JOURNAL/ NOT LOCAL AFTER JOURNAL T X X X DEFAULT JOURNAL TABLE T X X X CREATE FUNCTION A, T X X X Options RETURNS data_type/ RETURNS data_type CAST FROM data_type A X X X LANGUAGE C/ A X X X LANGUAGE CPP A X X X NO SQL A X X X SPECIFIC [database_name.] function_name A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 230 SQL Reference: Fundamentals CREATE FUNCTION, continued Options CLASS AGGREGATE/ CLASS AG T X X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X X DETERMINISTIC/ NOT DETERMINISTIC A X X X CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT A X X X EXTERNAL/ EXTERNAL NAME function_name/ EXTERNAL NAME function_name PARAMETER STYLE SQL/ EXTERNAL NAME function_name PARAMETER STYLE TD_GENERAL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE TD_GENERAL/ EXTERNAL NAME '[F delimiter function_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' A X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER A X X CREATE FUNCTION (table function form) T X X X Options RETURNS TABLE ( column_name data_type [ … , column_name data_type ] ) T X X X LANGUAGE C/ LANGUAGE CPP T X X X NO SQL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 231 CREATE FUNCTION (table function form), continued Options SPECIFIC [database_name.] function_name T X X X PARAMETER STYLE SQL T X X X DETERMINISTIC/ NOT DETERMINISTIC T X X X CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT T X X X EXTERNAL/ EXTERNAL NAME function_name/ EXTERNAL NAME function_name PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL NAME '[F delimiter function_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' T X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER T X X CREATE HASH INDEX T X X X Options FALLBACK PROTECTION/ NO FALLBACK PROTECTION T X X X ORDER BY VALUES/ ORDER BY HASH T X X X CHECKSUM = DEFAULT/ CHECKSUM = NONE/ CHECKSUM = LOW/ CHECKSUM = MEDIUM/ CHECKSUM = HIGH/ CHECKSUM = ALL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 232 SQL Reference: Fundamentals CREATE INDEX CREATE UNIQUE INDEX T X X X Options ALL T X X X ORDER BY VALUES/ ORDER BY HASH T X X X TEMPORARY T X X X CREATE JOIN INDEX T X X X Options FALLBACK PROTECTION/ NO FALLBACK PROTECTION T X X X CHECKSUM = DEFAULT/ CHECKSUM = NONE/ CHECKSUM = LOW/ CHECKSUM = MEDIUM/ CHECKSUM = HIGH/ CHECKSUM = ALL T X X X ROWID T X X X EXTRACT YEAR FROM/ EXTRACT MONTH FROM T X X X SUM numeric_expression T X X X COUNT column_expression T X X X FROM table_name/ FROM table_name correlation_name/ FROM table_name AS correlation_name T X X X FROM (joined_table) T X X X FROM table JOIN table/ FROM table INNER JOIN table/ FROM table LEFT JOIN table/ FROM table LEFT OUTER JOIN table/ FROM table RIGHT JOIN table/ FROM table RIGHT OUTER JOIN table T X X X |WHERE statement modifier| A X X X |GROUP BY statement modifier| T X X X |ORDER BY statement modifier| A, T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 233 CREATE JOIN INDEX, continued Options INDEX [index_name] [ALL] (column_list)/ INDEX [index_name] [ALL] (column_list) ORDER BY HASH [(column_name)]/ INDEX [index_name] [ALL] (column_list) ORDER BY VALUES [(column_name)]/ UNIQUE INDEX [index_name] (column_list)/ PRIMARY INDEX [index_name] (column_list)/ T X X X PRIMARY INDEX [index_name] (column_list) PARTITION BY expression T X CREATE MACRO/ CM T X X X Options AS statement T X X X USING modifier T X X X |LOCKING statement modifier| T X X X CREATE METHOD CREATE INSTANCE METHOD CREATE CONSTRUCTOR METHOD A X X Options EXTERNAL/ EXTERNAL NAME method_name/ EXTERNAL NAME '[F delimiter function_entry_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' A X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 234 SQL Reference: Fundamentals CREATE ORDERING A X X Options MAP WITH SPECIFIC METHOD specific_method_name/ MAP WITH METHOD method_name/ MAP WITH INSTANCE METHOD method_name/ MAP WITH SPECIFIC FUNCTION specific_function_name/ MAP WITH FUNCTION function_name A X X CREATE PROCEDURE (external stored procedure form) A X X X Options parameter_name data_type/ IN parameter_name data_type/ OUT parameter_name data_type/ INOUT parameter_name data_type A X X X LANGUAGE C/ LANGUAGE CPP A X X X NO SQL A X X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X X EXTERNAL/ EXTERNAL NAME procedure_name/ EXTERNAL NAME procedure_name PARAMETER STYLE SQL/ EXTERNAL NAME procedure_name PARAMETER STYLE TD_GENERAL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE TD_GENERAL/ EXTERNAL NAME '[F delimiter function_entry_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' A X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER A X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 235 CREATE PROCEDURE (stored procedure form) A, T X X X Options parameter_name data_type/ IN parameter_name data_type/ OUT parameter_name data_type/ INOUT parameter_name data_type A X X X NOT ATOMIC T X X X DECLARE variable-name data-type [DEFAULT literal] DECLARE variable-name data-type [DEFAULT NULL] A X X X DECLARE cursor_name [SCROLL] CURSOR FOR cursor_specification [FOR READ ONLY]/ DECLARE cursor_name [SCROLL] CURSOR FOR cursor_specification [FOR UPDATE]/ DECLARE cursor_name [NO SCROLL] CURSOR FOR cursor_specification [FOR READ ONLY]/ DECLARE cursor_name [NO SCROLL] CURSOR FOR cursor_specification [FOR UPDATE]/ A X X X DECLARE CONTINUE HANDLER DECLARE EXIT HANDLER A X X X FOR SQLSTATE sqlstate/ FOR SQLSTATE VALUE sqlstate A X X X FOR SQLEXCEPTION/ FOR SQLWARNING/ FOR NOT FOUND A X X X SET assignment_target = assignment_source A X X X IF expression THEN statement [ELSEIF expression THEN statement] [ELSE statement] END IF A X X X CASE operand1 WHEN operand2 THEN statement [ELSE statement] END CASE A X X X CASE WHEN expression THEN statement [ELSE statement] END CASE A X X X ITERATE label_name A X X X LEAVE label_name A X X X PRINT string_literal/ PRINT print_variable_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 236 SQL Reference: Fundamentals CREATE PROCEDURE, continued Options SQL_statement A X X X CALL procedure_name A X X X OPEN cursor_name A X X X CLOSE cursor_name A X X X FETCH [[NEXT] FROM] cursor_name INTO local_variable_name [ … , local_variable_name]/ FETCH [[FIRST] FROM] cursor_name INTO local_variable_name [ … , local_variable_name]/ FETCH [[NEXT] FROM] cursor_name INTO parameter_reference [ … , parameter_reference]/ FETCH [[FIRST] FROM] cursor_name INTO parameter_reference [ … , parameter_reference] A X X X WHILE expression DO statement END WHILE A X X X LOOP statement END LOOP A X X X FOR for_loop_variable AS [cursor_name CURSOR FOR] SELECT column_name [AS correlation_name] FROM table_name [WHERE clause] [SELECT clause] DO statement_list END FOR/ FOR for_loop_variable AS [cursor_name CURSOR FOR] SELECT expression [AS correlation_name] FROM table_name [WHERE clause] [SELECT clause] DO statement_list END FOR A X X X REPEAT statement_list UNTIL conditional_expression END REPEAT A X X X CREATE PROFILE T X X X Options ACCOUNT = ‘account_id’/ ACCOUNT = (‘account_id’ [ … ,’account_id’])/ ACCOUNT = NULL T X X X DEFAULT DATABASE = database_name/ DEFAULT DATABASE = NULL T X X X SPOOL = n [BYTES]/ SPOOL = NULL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 237 CREATE PROFILE, continued Options TEMPORARY = n [BYTES]/ TEMPORARY = NULL T X X X PASSWORD [ATTRIBUTES] = ( EXPIRE = n, EXPIRE = NULL, MINCHAR = n, MINCHAR = NULL, MAXCHAR = n, MAXCHAR = NULL, DIGITS = n, DIGITS = NULL, SPECCHAR = c, SPECCHAR = NULL, MAXLOGONATTEMPTS = n, MAXLOGONATTEMPTS = NULL, LOCKEDUSEREXPIRE = n, LOCKEDUSEREXPIRE = NULL, REUSE = n, REUSE = NULL) PASSWORD [ATTRIBUTES] = NULL T X X X CREATE REPLICATION GROUP A X X X CREATE ROLE A X X X CREATE TABLE/ CT A, T X X X Options SET/ MULTISET T X X X GLOBAL TEMPORARY A X X X GLOBAL TEMPORARY TRACE T X X X VOLATILE T X X X QUEUE T X X X FALLBACK [PROTECTION]/ NO FALLBACK [PROTECTION] T X X X WITH JOURNAL TABLE = name T X X X LOG/ NO LOG T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 238 SQL Reference: Fundamentals CREATE TABLE, continued Options [BEFORE] JOURNAL/ NO [BEFORE] JOURNAL/ DUAL [BEFORE] JOURNAL/ T X X X AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL JOURNAL/ NOT LOCAL JOURNAL T X X X FREESPACE = integer PERCENT T X X X DATABLOCKSIZE = integer/ DATABLOCKSIZE = integer BYTES/ DATABLOCKSIZE = integer KBYTES/ DATABLOCKSIZE = integer KILOBYTES T X X X MINIMUM DATABLOCKSIZE/ MAXIMUM DATABLOCKSIZE T X X X CHECKSUM = DEFAULT/ CHECKSUM = NONE/ CHECKSUM = LOW/ CHECKSUM = MEDIUM/ CHECKSUM = HIGH/ CHECKSUM = ALL T X X X QUEUE/ NO QUEUE T X X X column_name |Data Type| |Data Type Attributes| A X X X column_name |Data Type| |Column Storage Attributes| T X X X column_name |Data Type| |Column Constraint Attributes| A X X X GENERATED ALWAYS AS IDENTITY/ GENERATED BY DEFAULT AS IDENTITY A X X X |Column Constraint Attributes| T X X X |Table Constraint Attributes| T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 239 CREATE TABLE, continued Options [UNIQUE] [PRIMARY] INDEX [name] [ALL] (column_name) T X X X [UNIQUE] PRIMARY INDEX [name] (column) PARTITION BY expression T X X X INDEX [name] [ALL] (column_name) ORDER BY VALUES (name)/ INDEX [name] [ALL] (column_name) ORDER BY HASH (name) T X X X ON COMMIT DELETE ROWS/ ON COMMIT PRESERVE ROWS A X X X AS source_table_name WITH [NO] DATA/ A X X X AS source_table_name WITH [NO] DATA AND [NO] STATISTICS/ AS source_table_name WITH [NO] DATA AND [NO] STATS/ AS source_table_name WITH [NO] DATA AND [NO] STAT/ T X AS (query_expression) WITH [NO] DATA/ A X X X AS (query_expression) WITH [NO] DATA AND [NO] STATISTICS/ AS (query_expression) WITH [NO] DATA AND [NO] STATS/ AS (query_expression) WITH [NO] DATA AND [NO] STAT T X CREATE TRANSFORM A X X Options TO SQL WITH SPECIFIC METHOD specific_method_name/ TO SQL WITH METHOD method_name/ TO SQL WITH INSTANCE METHOD method_name/ TO SQL WITH SPECIFIC FUNCTION specific_function_name/ TO SQL WITH FUNCTION function_name A X X FROM SQL WITH SPECIFIC METHOD specific_method_name/ FROM SQL WITH METHOD method_name/ FROM SQL WITH INSTANCE METHOD method_name/ FROM SQL WITH SPECIFIC FUNCTION specific_function_name/ FROM SQL WITH FUNCTION function_name A X X CREATE TRIGGER A, T X X X Options ENABLED/ DISABLED T X X X BEFORE/ AFTER A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 240 SQL Reference: Fundamentals CREATE TRIGGER, continued Options INSERT ON table_name [ORDER integer]/ DELETE ON table_name [ORDER integer]/ UPDATE [OF (column_list)] ON table_name [ORDER integer] A X X X REFERENCING OLD_TABLE [AS] identifier [NEW_TABLE [AS] identifier]/ T X X X REFERENCING OLD [AS] identifier [NEW [AS] identifier]/ REFERENCING OLD TABLE [AS] identifier [NEW TABLE [AS] identifier]/ REFERENCING OLD [ROW] [AS] identifier [NEW [ROW] [AS] identifier] A X X X FOR EACH ROW/ FOR EACH STATEMENT A X X X WHEN (search_condition) A X X X (SQL_proc_statement ;)/ SQL_proc_statement / BEGIN ATOMIC (SQL_proc_statement;) END/ BEGIN ATOMIC SQL_proc_statement ; END A,T X X X CREATE TYPE (distinct form) A, T X X Options CHARACTER SET server_character_set T X X METHOD [SYSUDTLIB.]method_name/ INSTANCE METHOD [SYSUDTLIB.]method_name A, T X X RETURNS predefined_data_type/ RETURNS predefined_data_type AS LOCATOR/ RETURNS predefined_data_type [AS LOCATOR] CAST FROM predefined_data_type [AS LOCATOR]/ RETURNS predefined_data_type CAST FROM [SYSUDTLIB.]UDT_name [AS LOCATOR]/ RETURNS [SYSUDTLIB.]UDT_name/ RETURNS [SYSUDTLIB.]UDT_name AS LOCATOR/ RETURNS [SYSUDTLIB.]UDT_name [AS LOCATOR] CAST FROM predefined_data_type [AS LOCATOR]/ RETURNS [SYSUDTLIB.]UDT_name CAST FROM [SYSUDTLIB.]UDT_name [AS LOCATOR] A, T X X LANGUAGE C/ LANGUAGE CPP A X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 241 CREATE TYPE (distinct form), continued Options NO SQL A X X SPECIFIC [SYSUDTLIB.] specific_method_name A, T X X SELF AS RESULT A X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X DETERMINISTIC/ NOT DETERMINISTIC A X X CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT A X X CREATE TYPE (structured form) A, T X X Options AS (attribute_name predefined_data_type)/ AS (attribute_name predefined_data_type CHARACTER SET server_character_set)/ AS (attribute_name predefined_data_type [CHARACTER SET server_character_set] […, attribute_name predefined_data_type [CHARACTER SET server_character_set]] […, attribute_name UDT_name])/ AS (attribute_name predefined_data_type [CHARACTER SET server_character_set] […, attribute_name UDT_name] […, attribute_name predefined_data_type [CHARACTER SET server_character_set]])/ AS (attribute_name UDT_name)/ AS (attribute_name UDT_name […, attribute_name UDT_name] […, attribute_name predefined_data_type [CHARACTER SET server_character_set]])/ AS (attribute_name UDT_name […, attribute_name predefined_data_type [CHARACTER SET server_character_set]] […, attribute_name UDT_name]) A X X INSTANTIABLE A X X METHOD [SYSUDTLIB.]method_name/ INSTANCE METHOD [SYSUDTLIB.]method_name CONSTRUCTOR METHOD [SYSUDTLIB.]method_name A, T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 242 SQL Reference: Fundamentals CREATE TYPE (structured form), continued Options RETURNS predefined_data_type/ RETURNS predefined_data_type AS LOCATOR/ RETURNS predefined_data_type [AS LOCATOR] CAST FROM predefined_data_type [AS LOCATOR]/ RETURNS predefined_data_type CAST FROM [SYSUDTLIB.]UDT_name [AS LOCATOR]/ RETURNS [SYSUDTLIB.]UDT_name/ RETURNS [SYSUDTLIB.]UDT_name AS LOCATOR/ RETURNS [SYSUDTLIB.]UDT_name [AS LOCATOR] CAST FROM predefined_data_type [AS LOCATOR]/ RETURNS [SYSUDTLIB.]UDT_name CAST FROM [SYSUDTLIB.]UDT_name [AS LOCATOR] A, T X X LANGUAGE C/ LANGUAGE CPP A X X NO SQL A X X SPECIFIC [SYSUDTLIB.] specific_method_name A, T X X SELF AS RESULT A X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X DETERMINISTIC/ NOT DETERMINISTIC A X X CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT A X X CREATE USER T X X X Options FROM database_name T X X X PERMANENT = number [BYTES]/ PERM = number [BYTES] T X X X PASSWORD = password/ PASSWORD = NULL T X X X STARTUP = ‘string;’ T X X X TEMPORARY = n [bytes] T X X X SPOOL = n [BYTES] T X X X DEFAULT DATABASE = database_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 243 CREATE USER, continued Options COLLATION = collation_sequence T X X X ACCOUNT = ‘acct_ID’/ ACCOUNT = (‘acct_ID’ [ … ,’acct_ID’]) T X X X [NO] FALLBACK [PROTECTION] T X X X [BEFORE] JOURNAL/ NO [BEFORE] JOURNAL/ DUAL [BEFORE] JOURNAL T X X X AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL AFTER JOURNAL/ NOT LOCAL AFTER JOURNAL T X X X DEFAULT JOURNAL TABLE = table_name T X X X TIME ZONE = LOCAL/ TIME ZONE = [sign] quotestring/ TIME ZONE = NULL T X X X DATEFORM = INTEGERDATE/ DATEFORM = ANSIDATE T X X X DEFAULT CHARACTER SET data_type T X X X DEFAULT ROLE = role_name/ DEFAULT ROLE = NONE/ DEFAULT ROLE = NULL/ DEFAULT ROLE = ALL T X X X PROFILE = profile_name/ PROFILE = NULL T X X X CREATE VIEW A, T X X X Options (column_name [ … , column_name]) A X X X AS [ |LOCKING statement modifier| ] query_expression A, T X X X WITH CHECK OPTION A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 244 SQL Reference: Fundamentals CREATE RECURSIVE VIEW A X X X Options (column_name [ … , column_name]) A X X X AS (seed_statement [UNION ALL recursive_statement)] [ … [UNION ALL seed_statement] [ … UNION ALL recursive_statement]) A X X X DATABASE T X X X DECLARE CURSOR (selection form) A, T X X X Options FOR SELECT A X X X FOR COMMENT/ FOR EXPLAIN/ FOR HELP/ FOR SHOW T X X X DECLARE CURSOR (request form) A X X X Options FOR 'request_specification' A X X X DECLARE CURSOR (macro form) T X X X Options FOR EXEC macro_name T X X X DECLARE CURSOR (dynamic SQL form) A X X X Options FOR statement_name A X X X DECLARE STATEMENT T X X X DECLARE TABLE T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 245 DELETE (basic/searched form)/ DEL A, T X X X Options [FROM] table_name A X X X [AS] alias_name A X X X WHERE condition A X X X ALL T X X X DELETE (implied join condition form)/ DEL A, T X X X Options delete_table_name T X X X [FROM] table_name [ … ,[FROM] table_name] T X X X [AS] alias_name A X X X WHERE condition A X X X ALL T X X X DELETE (positioned form)/ DEL A X X X Options FROM table_name A X X X WHERE CURRENT OF cursor_name A X X X DELETE DATABASE DELETE USER T X X X Option ALL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 246 SQL Reference: Fundamentals DESCRIBE T X X X Options INTO descriptor_area T X X X USING NAMES/ USING ANY/ USING BOTH/ USING LABELS T X X X FOR STATEMENT statement_number/ FOR STATEMENT [:] num_var T X X X DIAGNOSTIC "validate index" T X X X Option ON/ NOT ON T X X X DIAGNOSTIC DUMP SAMPLES T X X X DIAGNOSTIC HELP SAMPLES T X X X DIAGNOSTIC SET SAMPLES T X X X Options ON/ NOT ON T X X X FOR SESSION/ FOR SYSTEM T X X X DROP AUTHORIZATION T X X DROP CAST A X X DROP DATABASE DROP USER T X X X DROP FUNCTION DROP SPECIFIC FUNCTION A X X X DROP HASH INDEX T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 247 DROP INDEX T X X X Options TEMPORARY T X X X ORDER BY (column_name)/ ORDER BY VALUES (column_name)/ ORDER BY HASH (column_name) T X X X DROP JOIN INDEX T X X X DROP MACRO T X X X DROP ORDERING A X X DROP PROCEDURE A X X X DROP PROFILE T X X X DROP REPLICATION GROUP T X X X DROP ROLE A X X X DROP STATISTICS/ DROP STATS/ DROP STAT (optimizer form) T X X X Options [FOR] [UNIQUE] INDEX index_name/ [FOR] [UNIQUE] INDEX [index_name] (col_name) [ORDER BY col_name]/ [FOR] [UNIQUE] INDEX [index_name] (col_name) [ORDER BY VALUES (col_name)]/ [FOR] [UNIQUE] INDEX [index_name] (col_name) [ORDER BY HASH (col_name)]/ [FOR] COLUMN column_name/ [FOR] COLUMN (column_name [ … , column_name])/ T X X X [FOR] COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ [FOR] COLUMN (PARTITION [ … , column_name])/ [FOR] COLUMN PARTITION T X X ON T X X X TEMPORARY T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 248 SQL Reference: Fundamentals Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0 DROP STATISTICS/ DROP STATS/ DROP STAT (QCD form) T X X X Options INDEX (column_name [ … , column_name])/ INDEX index_name/ COLUMN (column_name [ … ,column_name])/ COLUMN column_name/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X DROP TABLE A, T X X X Options TEMPORARY A X X X ALL A X X X OVERRIDE A X X X DROP TRANSFORM A X X DROP TRIGGER T X X X DROP TYPE A X X DROP VIEW T X X X DUMP EXPLAIN T X X X Options AS query_plan_name T X X X LIMIT/ LIMIT SQL/ LIMIT SQL = n T X X X CHECK STATISTICS T X X ECHO T X X X END DECLARE SECTION T X X X END-EXEC A X X XAppendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 249 END LOGGING T X X X Options DENIALS T X X X WITH TEXT T X X X ALL/ operation/ GRANT T X X X BY database_name T X X X ON DATABASE name/ ON FUNCTION/ ON MACRO name/ ON PROCEDURE name/ ON TABLE name/ ON TRIGGER name/ ON USER name/ ON VIEW name T X X X END QUERY LOGGING T X X X Options ON ALL/ ON user_name/ ON user_name ACCOUNT = 'account_name'/ ON user_name ACCOUNT = ('account_name' [ … ,'account_name']) T X X X END TRANSACTION/ ET T X X X EXECUTE macro_name/ EXEC macro_name T X X X EXECUTE statement_name A X X X Options USING [:] host_variable_name A X X X [INDICATOR] :host_indicator_name A X X X USING DESCRIPTOR [:] descriptor_area A X X X EXECUTE IMMEDIATE A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 250 SQL Reference: Fundamentals FETCH A X X X Options INTO [:] host_variable_name A X X X [INDICATOR] :host_indicator_name A X X X USING DESCRIPTOR [:] descriptor_area A X X X GET CRASH (embedded SQL) T X X X GIVE T X X X Options database_name TO recipient_name/ user_name TO recipient_name T X X X GRANT A, T X X X Options ALL/ A X X X ALL PRIVILEGES/ ALL BUT T X X X DELETE/ EXECUTE/ INSERT/ REFERENCES/ SELECT/ UPDATE/ A X X X ALTER/ CHECKPOINT/ CREATE/ DROP/ DUMP/ INDEX/ RESTORE/ T X X X REPLCONTROL/ T X X X UDTMETHOD/ UDTTYPE/ UDTUSAGE T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 251 GRANT, continued Options ON database_name/ ON database_name.object_name/ ON object_name/ ON PROCEDURE identifier/ ON SPECIFIC FUNCTION specific_function_name/ ON FUNCTION function_name/ A X X X ON TYPE UDT_name/ ON TYPE SYSUDTLIB.UDT_name A X X TO user_name/ TO ALL user_name/ T X X X TO PUBLIC A X X X WITH GRANT OPTION A X X X GRANT LOGON T X X X Options ON host_id/ ON ALL T X X X AS DEFAULT/ TO database_name/ FROM database_name T X X X WITH NULL PASSWORD T X X X GRANT MONITOR/ GRANT monitor_privilege T X X X Options PRIVILEGES/ BUT NOT monitor_privilege T X X X TO [ALL] user_name/ TO PUBLIC T X X X WITH GRANT OPTION T X X X GRANT ROLE A X X X Options WITH ADMIN OPTION A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 252 SQL Reference: Fundamentals HELP T X X X Options CAST [database_name.] UDT_name/ CAST [database_name.] UDT_name SOURCE/ CAST [database_name.] UDT_name TARGET T X X COLUMN column_name FROM table_name/ COLUMN * FROM table_name/ COLUMN table_name.column_name/ COLUMN table_name.*/ COLUMN expression T X X X CONSTRAINT [database_name.] table_name.name T X X X DATABASE database_name T X X X FUNCTION function_name [(data_type [ … , data_type])]/ SPECIFIC FUNCTION specific_function_name T X X X HASH INDEX hash_index_name T X X X [TEMPORARY] INDEX table_name [(column_name)]/ [TEMPORARY] INDEX join_index_name [(column_name)] T X X X JOIN INDEX join_index_name T X X X MACRO macro_name T X X X METHOD [database_name.] method_name/ INSTANCE METHOD [database_name.] method_name/ CONSTRUCTOR METHOD [database_name.] method_name/ SPECIFIC METHOD [database_name.] specific_method_name T X X PROCEDURE [database_name.] procedure_name/ PROCEDURE [database_name.] procedure_name ATTRIBUTES/ PROCEDURE [database_name.] procedure_name ATTR/ PROCEDURE [database_name.] procedure_name ATTRS T X X X REPLICATION GROUP T X X X SESSION T X X X TABLE table_name/ TABLE join_index_name T X X X TRANSFORM [database_name.] UDT_name T X X TRIGGER [database_name.] trigger_name/ TRIGGER [database_name.] table_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 253 HELP, continued Options TYPE [database_name.] UDT_name/ TYPE [database_name.] UDT_name ATTRIBUTE/ TYPE [database_name.] UDT_name METHOD T X X USER user_name T X X X VIEW view_name T X X X VOLATILE TABLE T X X X HELP STATISTICS/ HELP STATS/ HELP STAT (optimizer form) T X X X Option INDEX (column_name [ … , column_name])/ INDEX index_name/ COLUMN (column_name [ … ,column_name])/ COLUMN column_name/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X HELP STATISTICS/ HELP STATS/ HELP STAT (QCD form) T X X X Options INDEX (column_name [ … , column_name])/ INDEX index_name/ COLUMN (column_name [ … ,column_name])/ COLUMN column_name/ T X X X COLUMN (column_name [ … , column_name], PARTITION [ … , column_name])/ COLUMN (PARTITION [ … , column_name])/ COLUMN PARTITION T X X FOR QUERY query_ID T X X SAMPLEID statistics_ID T X X UPDATE MODIFIED T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 254 SQL Reference: Fundamentals INCLUDE A X X X INCLUDE SQLCA T X X X INCLUDE SQLDA T X X X INITIATE INDEX ANALYSIS T X X X Options ON table_name [ … , table_name] T X X X SET IndexesPerTable = value [, SearchSpace = value] [, ChangeRate = value] [, ColumnsPerIndex = value] T X X X [, JoinIndexesPerTable = value] [, ColumnsPerJoinIndex = value] [, IndexMaintMode = value] T X X KEEP INDEX T X X X USE MODIFIED STATISTICS/ USE MODIFIED STATS/ USE MODIFIED STAT T X X X WITH INDEX TYPE number/ WITH INDEX TYPE number [ … , number]/ WITH NO INDEX TYPE number/ WITH NO INDEX TYPE number [ … , number] T X X X CHECKPOINT checkpoint_trigger T X X X INSERT/ INS A T X X X Options [VALUES] (expression [ … , expression]) A X X X (column_name [ … , column_name]) VALUES (expression [ … , expression]) A X X X [(column_name [ … , column_name])] subquery A X X X DEFAULT VALUES A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 255 INSERT EXPLAIN T X X X Options WITH [NO] STATISTICS T X X X AND DEMOGRAPHICS T X X X USING SAMPLE percentage/ USING SAMPLE percentage PERCENT T X X FOR table_name [ … , table_name] T X X X AS query_plan_name T X X X LIMIT/ LIMIT SQL/ LIMIT SQL = n T X X X FOR frequency T X X X LOGOFF (embedded SQL) T X X X Options CURRENT/ ALL/ connection_name/ :host_variable_name T X X X LOGON (embedded SQL) T X X X Options AS connection_name/ AS :namevar T X X X MERGE A X X X Options INTO A X X X AS correlation_name A X X X VALUES using_expression/ (subquery) A X X X ON match_condition A X X X WHEN MATCHED THEN UPDATE SET/ WHEN NOT MATCHED THEN INSERT A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 256 SQL Reference: Fundamentals MODIFY DATABASE T X X X Options PERMANENT = number [BYTES]/ PERM = number [BYTES] T X X X TEMPORARY = number [bytes] T X X X SPOOL = number [BYTES] T X X X ACCOUNT = ‘account_ID’ T X X X [NO] FALLBACK [PROTECTION] T X X X [BEFORE] JOURNAL/ NO [BEFORE] JOURNAL/ DUAL [BEFORE] JOURNAL T X X X AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL AFTER JOURNAL/ NOT LOCAL AFTER JOURNAL T X X X DEFAULT JOURNAL TABLE = table_name T X X X DROP DEFAULT JOURNAL TABLE [= table_name] T X X X MODIFY PROFILE T X X X Options ACCOUNT = ‘account_id’/ ACCOUNT = (‘account_id’ [ … ,’account_id’])/ ACCOUNT = NULL T X X X DEFAULT DATABASE = database_name/ DEFAULT DATABASE = NULL T X X X SPOOL = n [BYTES]/ SPOOL = NULL T X X X TEMPORARY = n [BYTES]/ TEMPORARY = NULL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 257 MODIFY PROFILE, continued Options PASSWORD [ATTRIBUTES] = ( EXPIRE = n, EXPIRE = NULL, MINCHAR = n, MINCHAR = NULL, MAXCHAR = n, MAXCHAR = NULL, DIGITS = n, DIGITS = NULL, SPECCHAR = c, SPECCHAR = NULL, MAXLOGONATTEMPTS = n, MAXLOGONATTEMPTS = NULL, LOCKEDUSEREXPIRE = n, LOCKEDUSEREXPIRE = NULL, REUSE = n, REUSE = NULL) PASSWORD [ATTRIBUTES] = NULL T X X X MODIFY USER T X X X Options PERMANENT = number [BYTES]/ PERM = number [BYTES] T X X X PASSWORD = password [FOR USER] T X X X STARTUP = ‘string;’/ STARTUP = NULL T X X X RELEASE PASSWORD LOCK T X X X TEMPORARY = n [bytes] T X X X SPOOL = n [BYTES] T X X X ACCOUNT = ‘acct_ID’ ACCOUNT = (‘acct_ID’ [ … ,’acct_ID’]) T X X X DEFAULT DATABASE = database_name T X X X COLLATION = collation_sequence T X X X [NO] FALLBACK [PROTECTION] T X X X [BEFORE] JOURNAL/ NO [BEFORE] JOURNAL/ DUAL [BEFORE] JOURNAL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 258 SQL Reference: Fundamentals MODIFY USER, continued Options AFTER JOURNAL/ NO AFTER JOURNAL/ DUAL AFTER JOURNAL/ LOCAL AFTER JOURNAL/ NOT LOCAL AFTER JOURNAL T X X X DEFAULT JOURNAL TABLE = table_name T X X X DROP DEFAULT JOURNAL TABLE [= table_name] T X X X TIME ZONE = LOCAL/ TIME ZONE = [sign] quotestring/ TIME ZONE = NULL T X X X DATEFORM = INTEGERDATE/ DATEFORM = ANSIDATE T X X X DEFAULT CHARACTER SET data_type T X X X DEFAULT ROLE T X X X PROFILE T X X X OPEN A X X X Options USING [:] host_variable_name A X X X [INDICATOR] :host_indicator_name A X X X USING DESCRIPTOR [:] descriptor_area A X X X POSITION A X X X Options TO NEXT/ TO [STATEMENT] statement_number/ TO [STATEMENT] [:] numvar A X X X PREPARE A X X X Options INTO [:] descriptor_area A X X X USING NAMES/ USING ANY/ USING BOTH/ USING LABELS A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 259 PREPARE, continued Options FOR STATEMENT statement_number/ FOR STATEMENT [:] numvar A X X X FROM statement_string/ FROM [:] statement_string_var A X X X RENAME FUNCTION T X X X RENAME MACRO T X X X RENAME PROCEDURE T X X X RENAME TABLE T X X X RENAME TRIGGER T X X X RENAME VIEW T X X X REPLACE CAST T X X Options WITH SPECIFIC METHOD specific_method_name/ WITH METHOD method_name/ WITH INSTANCE METHOD method_name/ WITH SPECIFIC FUNCTION specific_function_name/ WITH FUNCTION function_name T X X AS ASSIGNMENT T X X REPLACE FUNCTION T X X X Options RETURNS data_type/ RETURNS data_type CAST FROM data_type A X X X LANGUAGE C/ LANGUAGE CPP A X X X NO SQL A X X X SPECIFIC [database_name.] function_name A X X X CLASS AGGREGATE/ CLASS AG T X X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X X DETERMINISTIC/ NOT DETERMINISTIC A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 260 SQL Reference: Fundamentals REPLACE FUNCTION, continued Options CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT A X X X EXTERNAL/ EXTERNAL NAME function_name/ EXTERNAL NAME function_name PARAMETER STYLE SQL/ EXTERNAL NAME function_name PARAMETER STYLE TD_GENERAL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE TD_GENERAL/ EXTERNAL NAME '[F delimiter function_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' A X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER A X X REPLACE FUNCTION (table function form) T X X X Options RETURNS TABLE ( column_name data_type [ … , column_name data_type ] ) T X X X LANGUAGE C/ LANGUAGE CPP T X X X NO SQL T X X X SPECIFIC [database_name.] function_name T X X X PARAMETER STYLE SQL T X X X DETERMINISTIC/ NOT DETERMINISTIC T X X X CALLED ON NULL INPUT/ RETURNS NULL ON NULL INPUT T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 261 REPLACE FUNCTION (table function form), continued Options EXTERNAL/ EXTERNAL NAME function_name/ EXTERNAL NAME function_name PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL NAME '[F delimiter function_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' T X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER T X X REPLACE MACRO T X X X Options AS T X X X USING T X X X REPLACE METHOD REPLACE CONSTRUCTOR METHOD REPLACE INSTANCE METHOD REPLACE SPECIFIC METHOD T X X Options parameter_name data_type/ parameter_name UDT_name T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 262 SQL Reference: Fundamentals REPLACE METHOD, continued Options EXTERNAL/ EXTERNAL NAME method_name/ EXTERNAL NAME '[F delimiter function_entry_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' T X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER T X X REPLACE ORDERING A X X Options MAP WITH SPECIFIC METHOD specific_method_name/ MAP WITH METHOD method_name/ MAP WITH INSTANCE METHOD method_name/ MAP WITH SPECIFIC FUNCTION specific_function_name/ MAP WITH FUNCTION function_name A X X REPLACE PROCEDURE (external stored procedure form) A X X X Options parameter_name data_type/ IN parameter_name data_type/ OUT parameter_name data_type/ INOUT parameter_name data_type A X X X LANGUAGE C/ LANGUAGE CPP A X X X NO SQL A X X X PARAMETER STYLE SQL/ PARAMETER STYLE TD_GENERAL A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 263 REPLACE PROCEDURE (external stored procedure form), continued Options EXTERNAL/ EXTERNAL NAME procedure_name/ EXTERNAL NAME procedure_name PARAMETER STYLE SQL/ EXTERNAL NAME procedure_name PARAMETER STYLE TD_GENERAL/ EXTERNAL PARAMETER STYLE SQL/ EXTERNAL PARAMETER STYLE TD_GENERAL/ EXTERNAL NAME '[F delimiter function_entry_name] [D] [SI delimiter name delimiter include_name] [CI delimiter name delimiter include_name] [SL delimiter library_name] [SO delimiter name delimiter object_name ] [CO delimiter name delimiter object_name] [SP delimiter package_name] [SS delimiter name delimiter source_name] [CS delimiter name delimiter source_name]' A X X X EXTERNAL SECURITY DEFINER/ EXTERNAL SECURITY DEFINER authorization_name/ EXTERNAL SECURITY INVOKER A X X REPLACE PROCEDURE (stored procedure form) T X X X Options parameter_name data_type/ IN parameter_name data_type/ OUT parameter_name data_type/ INOUT parameter_name data_type T X X X NOT ATOMIC T X X X DECLARE variable-name data-type [DEFAULT literal] DECLARE variable-name data-type [DEFAULT NULL] T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 264 SQL Reference: Fundamentals REPLACE PROCEDURE (stored procedure form), continued Options DECLARE cursor_name [SCROLL] CURSOR FOR cursor_specification [FOR READ ONLY]/ DECLARE cursor_name [SCROLL] CURSOR FOR cursor_specification [FOR UPDATE]/ DECLARE cursor_name [NO SCROLL] CURSOR FOR cursor_specification [FOR READ ONLY]/ DECLARE cursor_name [NO SCROLL] CURSOR FOR cursor_specification [FOR UPDATE]/ T X X X DECLARE CONTINUE HANDLER/ DECLARE EXIT HANDLER T X X X FOR SQLSTATE sqlstate/ FOR SQLSTATE VALUE sqlstate T X X X FOR SQLEXCEPTION/ FOR SQLWARNING/ FOR NOT FOUND T X X X SET assignment_target = assignment_source T X X X IF expression THEN statement [ELSEIF expression THEN statement] [ELSE statement] END IF T X X X CASE operand1 WHEN operand2 THEN statement [ELSE statement] END CASE T X X X CASE WHEN expression THEN statement [ELSE statement] END CASE T X X X ITERATE label_name T X X X LEAVE label_name T X X X PRINT string_literal/ PRINT print_variable_name T X X X SQL_statement T X X X CALL procedure_name T X X X OPEN cursor_name T X X X CLOSE cursor_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 265 REPLACE PROCEDURE (stored procedure form), continued Options FETCH [[NEXT] FROM] cursor_name INTO local_variable_name [ … , local_variable_name]/ FETCH [[FIRST] FROM] cursor_name INTO local_variable_name [ … , local_variable_name]/ FETCH [[NEXT] FROM] cursor_name INTO parameter_reference [ … , parameter_reference]/ FETCH [[FIRST] FROM] cursor_name INTO parameter_reference [ … , parameter_reference] T X X X WHILE expression DO statement END WHILE T X X X LOOP statement END LOOP T X X X FOR for_loop_variable AS [cursor_name CURSOR FOR] SELECT column_name [AS correlation_name] FROM table_name [WHERE clause] [SELECT clause] DO statement_list END FOR/ FOR for_loop_variable AS [cursor_name CURSOR FOR] SELECT expression [AS correlation_name] FROM table_name [WHERE clause] [SELECT clause] DO statement_list END FOR T X X X REPEAT statement_list UNTIL conditional_expression END REPEAT T X X X REPLACE TRANSFORM T X X Options TO SQL WITH SPECIFIC METHODspecific_method_name/ TO SQL WITH METHOD method_name/ TO SQL WITH INSTANCE METHOD method_name/ TO SQL WITH SPECIFIC FUNCTION specific_function_name/ TO SQL WITH FUNCTION function_name T X X FROM SQL WITH SPECIFIC METHOD specific_method_name/ FROM SQL WITH METHOD method_name/ FROM SQL WITH INSTANCE METHOD method_name/ FROM SQL WITH SPECIFIC FUNCTION specific_function_name/ FROM SQL WITH FUNCTION function_name T X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 266 SQL Reference: Fundamentals REPLACE TRIGGER T X X X Options ENABLED/ DISABLED T X X X BEFORE/ AFTER T X X X INSERT/ DELETE/ UPDATE [OF (column_list)] T X X X ORDER integer T X X X REFERENCING OLD_TABLE [AS] identifier [NEW_TABLE [AS] identifier]/ REFERENCING OLD [AS] identifier [NEW [AS] identifier]/ REFERENCING OLD TABLE [AS] identifier [NEW TABLE [AS] identifier]/ REFERENCING OLD [ROW] [AS] identifier [NEW [ROW] [AS] identifier] T X X X FOR EACH ROW/ FOR EACH STATEMENT T X X X WHEN (search_condition) T X X X (SQL_proc_statement ;)/ SQL_proc_statement / BEGIN ATOMIC (SQL_proc_statement;) END/ BEGIN ATOMIC SQL_proc_statement ; END T X X X REPLACE VIEW A, T X X X Options (column_name [ … , column_name]) T X X X AS [ |LOCKING statement modifier| ] query_expression A, T X X X WITH CHECK OPTION A X X X RESTART INDEX ANALYSIS T X X X REVOKE A, T X X X Options GRANT OPTION FOR A X X X ALL/ ALL PRIVILEGES/ ALL BUT operation A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 267 REVOKE, continued Options DELETE/ INSERT/ SELECT/ REFERENCES/ UPDATE/ A X X X ALTER/ CHECKPOINT/ CREATE/ DROP/ DUMP/ EXECUTE/ INDEX/ RESTORE/ T X X X REPLCONTROL/ T X X X UDTMETHOD/ UDTTYPE/ UDTUSAGE T X X ON database_name/ ON database_name.object_name/ ON object_name/ ON PROCEDURE procedure_name/ ON SPECIFIC FUNCTION specific_function_name/ ON FUNCTION function_name/ A X X X ON TYPE UDT_name/ ON TYPE SYSUDTLIB.UDT_name A X X TO [ALL] user_name/ TO PUBLIC/ FROM [ALL] user_name/ FROM PUBLIC T X X X REVOKE LOGON T X X X Options ON host_id/ ON ALL T X X X AS DEFAULT/ TO database_name/ FROM database_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 268 SQL Reference: Fundamentals REVOKE MONITOR/ REVOKE monitor_privilege T X X X Options GRANT OPTION FOR T X X X PRIVILEGES/ BUT NOT monitor_privilege T X X X TO [ALL] user_name/ TO PUBLIC/ FROM [ALL] user_name/ FROM PUBLIC T X X X REVOKE ROLE A X X X Options ADMIN OPTION FOR A X X X REWIND T X X X ROLLBACK A, T X X X Options WORK A X X X WORK RELEASE T X X X 'abort_message' T X X X FROM_clause T X X X WHERE_clause T X X X SELECT/ SEL A, T X X X Options |WITH [RECURSIVE] statement modifier| A X X X DISTINCT/ ALL A X X X TOP integer [WITH TIES]/ TOP integer PERCENT [WITH TIES]/ TOP decimal [WITH TIES]/ TOP decimal PERCENT [WITH TIES] T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 269 SELECT, continued Options */ expression/ expression [AS] alias_name/ table_name.*/ A X X X *.ALL/ table_name.*.ALL/ column_name.ALL T X X SAMPLEID T X X X FROM table_name/ FROM table_name [AS] alias_name/ FROM join_table_name JOIN joined_table ON search_condition/ FROM join_table_name INNER JOIN joined_table ON search_condition/ FROM join_table_name LEFT JOIN joined_table ON search_condition/ FROM join_table_name LEFT OUTER JOIN joined_table ON search_condition/ FROM join_table_name RIGHT JOIN joined_table ON search_condition/ FROM join_table_name RIGHT OUTER JOIN joined_table ON search_condition/ FROM join_table_name FULL JOIN joined_table ON search_condition/ FROM join_table_name FULL OUTER JOIN joined_table ON search_condition/ FROM join_table_name CROSS JOIN/ FROM (subquery) [AS] derived_table_name/ FROM (subquery) [AS] derived_table_name (column_name)/ FROM TABLE (function_name([expression [ … , expression]])) [AS] derived_table_name/ FROM TABLE (function_name([expression [ … , expression]])) [AS] derived_table_name (column_name [ … , column_name]) A X X X |WHERE statement modifier| A X X X |GROUP BY statement modifier| A, T X X X |HAVING statement modifier| A X X X |QUALIFY statement modifier| T X X X |SAMPLE statement modifier| T X X X |ORDER BY statement modifier| A, T X X X |WITH statement modifier| T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 270 SQL Reference: Fundamentals SELECT AND CONSUME TOP 1 T X X X Options FROM queue_table_name T X X X SELECT … INTO/ SEL … INTO A, T X X X Options DISTINCT/ ALL A X X X AND CONSUME TOP 1 T X X X expression/ expression [AS] alias_name A X X X FROM table_name/ FROM table_name [AS] alias_name/ FROM join_table_name JOIN joined_table ON search_condition/ FROM join_table_name INNER JOIN joined_table ON search_condition/ FROM join_table_name LEFT JOIN joined_table ON search_condition/ FROM join_table_name LEFT OUTER JOIN joined_table ON search_condition/ FROM join_table_name RIGHT JOIN joined_table ON search_condition/ FROM join_table_name RIGHT OUTER JOIN joined_table ON search_condition/ FROM join_table_name FULL JOIN joined_table ON search_condition/ FROM join_table_name FULL OUTER JOIN joined_table ON search_condition/ FROM join_table_name CROSS JOIN/ FROM (subquery) [AS] derived_table_name/ FROM (subquery) [AS] derived_table_name (column_name) A X X X |WHERE statement modifier| A X X X SET BUFFERSIZE (embedded SQL) T X X X SET CHARSET (embedded SQL) T X X X SET CONNECTION (embedded SQL) T X X X SET CRASH (embedded SQL) T X X X Options WAIT_NOTELL/ NOWAIT_TELL T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 271 SET ROLE A, T X X X Options role_name/ NONE/ A X X X NULL/ ALL/ EXTERNAL T X X X SET SESSION ACCOUNT/ SS ACCOUNT T X X X Options FOR SESSION/ FOR REQUEST T X X X SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL/ SS CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL A X X Options RU/ READ UNCOMMITTED/ SR/ SERIALIZABLE A X X SET SESSION COLLATION/ SS COLLATION T X X X SET SESSION DATABASE/ SS DATABASE T X X X SET SESSION DATEFORM/ SS DATEFORM T X X X Options ANSIDATE/ INTEGERDATE T X X X SET SESSION FUNCTION TRACE/ SS FUNCTION TRACE T X X X Options OFF/ USING mask FOR TABLE table_name/ USING mask FOR TRACE TABLE table_name T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 272 SQL Reference: Fundamentals SET SESSION OVERRIDE REPLICATION/ SS OVERRIDE REPLICATION T X X X Options OFF/ ON T X X X SET TIME ZONE T X X X Options LOCAL/ INTERVAL offset HOUR TO MINUTE/ USER T X X X SHOW T X X X Options QUALIFIED T X X X SHOW CAST T X SHOW FUNCTION SHOW SPECIFIC FUNCTION T X X X SHOW HASH INDEX T X X X SHOW JOIN INDEX T X X X SHOW MACRO T X X X SHOW METHOD SHOW CONSTRUCTOR METHOD SHOW INSTANCE METHOD SHOW SPECIFIC METHOD T X X SHOW PROCEDURE T X X X SHOW REPLICATION GROUP T X X X SHOW [TEMPORARY] TABLE T X X X SHOW TRIGGER T X X X SHOW TYPE T X X SHOW VIEW T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 273 TEST T X X X Options async_statement_identifier/ :namevar T X X X COMPLETION T X X X UPDATE/ UPD (searched form) A, T X X X Options table_name A X X X [AS] alias_name/ FROM table_name [[AS] alias_name] [ … , table_name [[AS] alias_name]] A, T X X X SET column_name=expression [ … , column_name=expression]/ A X X X SET column_name=expression [ … , column_name=expression] [ … , column_name.mutator_name=expression]/ SET column_name.mutator_name=expression [ … , column_name.mutator_name=expression] [ … , column_name=expression] A X X ALL T X X X |WHERE statement modifier| A X X X UPDATE/ UPD (positioned form) A X X X Options table_name [alias_name] A X X X SET column_name=expression [ … , column_name=expression] A X X X WHERE CURRENT OF cursor_name A X X X UPDATE/ UPD (upsert form) T X X X Options table_name_1 T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 274 SQL Reference: Fundamentals UPDATE (upsert form), continued Options SET column_name=expression [ … , column_name=expression]/ T X X X SET column_name=expression [ … , column_name=expression] [ … , column_name.mutator_name=expression]/ SET column_name.mutator_name=expression [ … , column_name.mutator_name=expression] [ … , column_name=expression] T X X |WHERE statement modifier| T X X X ELSE INSERT [INTO] table_name_2/ ELSE INS [INTO] table_name_2 T X X X [(column_name [ … , column_name])] VALUES (expression)/ DEFAULT VALUES T X X X WAIT T X X X Options async_statement_identifier COMPLETION/ ALL COMPLETION/ ANY COMPLETION INTO [:] stmtvar, [:] sessvar T X X X WHENEVER A, T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers SQL Reference: Fundamentals 275 Request Modifier EXPLAIN T X X X Statement Modifiers ASYNC T X X X EXEC SQL A X X X GROUP BY clause A, T X X X Options CUBE/ GROUPING SETS/ ROLLUP A X X X HAVING clause A X X X LOCKING/ LOCK T X X X Options DATABASE database_name/ TABLE table_name/ VIEW view_name/ ROW T X X X FOR/ IN T X X X ACCESS/ EXCLUSIVE/ EXCL/ SHARE/ WRITE/ CHECKSUM/ READ/ READ OVERRIDE T X X X MODE T X X X NOWAIT T X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Statements and Modifiers 276 SQL Reference: Fundamentals Statement Modifiers, continued ORDER BY clause A, T X X X Options expression T X X X column_name/ column_position A X X X ASC/ DESC A X X X QUALIFY clause T X X X SAMPLE clause T X X X Options WITH REPLACEMENT T X X X RANDOMIZED ALLOCATION T X X X USING row descriptor T X X X Options AS DEFERRED/ AS LOCATOR T X X X WHERE clause A X X X WITH clause T X X X Options expression_1 T X X X BY expression_2 T X X X ASC/ DESC T X X X WITH [RECURSIVE] clause A X X X Options (column_name [ … , column_name]) A X X X AS (seed_statement [UNION ALL recursive_statement)] [ … [UNION ALL seed_statement] [ … UNION ALL recursive_statement]) A X X X Statement ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Data Types and Literals SQL Reference: Fundamentals 277 Data Types and Literals The following list contains all SQL data types and literals for this version and previous versions of Teradata Database. The following type codes appear in the ANSI Compliance column. Code Definition A ANSI T Teradata extension Data Type / Literal ANSI Compliance V2R6.2 V2R6.1 V2R6.0 Data Types BIGINT A X BINARY LARGE OBJECT, BLOB A X X X BYTE T X X X BYTEINT T X X X CHAR, CHARACTER A X X X CHAR VARYING, CHARACTER VARYING A X X X CHARACTER LARGE OBJECT, CLOB A X X X DATE A, T X X X DEC, DECIMAL A X X X DOUBLE PRECISION A X X X FLOAT A X X X GRAPHIC T X X X INT, INTEGER A X X X INTERVAL DAY A X X X INTERVAL DAY TO HOUR A X X X INTERVAL DAY TO MINUTE A X X X INTERVAL DAY TO SECOND A X X X INTERVAL HOUR A X X X INTERVAL HOUR TO MINUTE A X X XAppendix E: SQL Feature Summary Data Types and Literals 278 SQL Reference: Fundamentals Data Types, continued INTERVAL HOUR TO SECOND A X X X INTERVAL MINUTE A X X X INTERVAL MINUTE TO SECOND A X X X INTERVAL MONTH A X X X INTERVAL SECOND A X X X INTERVAL YEAR A X X X INTERVAL YEAR TO MONTH A X X X LONG VARCHAR T X X X LONG VARGRAPHIC T X X X NUMERIC A X X X REAL A X X X SMALLINT A X X X TIME A X X X TIME WITH TIMEZONE A X X X TIMESTAMP A X X X TIMESTAMP WITH TIMEZONE A X X X user-defined type (UDT) A X X VARBYTE T X X X VARCHAR A X X X VARGRAPHIC T X X X Literals Character data A X X X DATE A X X X Decimal A X X X Floating point A X X X Graphic T X X X Hexadecimal T X X X Integer A X X X Data Type / Literal ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Data Types and Literals SQL Reference: Fundamentals 279 Literals, continued Interval A X X X TIME A X X X TIMESTAMP A X X X Data Type Attributes AS output format phrase A X X X CASESPECIFIC/NOT CASESPECIFIC phrase/ CS/NOT CS phrase T X X X CHARACTER SET A X X X CHECK table constraint attribute A X X X COMPRESS/ COMPRESS NULL/ COMPRESS string/ COMPRESS value column storage attribute T X X X COMPRESS (value_list) column storage attribute T X X X CONSTRAINT/ CONSTRAINT CHECK/ CONSTRAINT PRIMARY KEY/ CONSTRAINT REFERENCES/ CONSTRAINT UNIQUE column constraint attribute T X X X DEFAULT constant_value/ DEFAULT DATE quotestring/ DEFAULT INTERVAL quotestring/ DEFAULT TIME quotestring/ DEFAULT TIMESTAMP quotestring default value control phrase A X X X FOREIGN KEY table constraint attribute A X X X FORMAT output format phrase T X X X NAMED output format phrase T X X X NOT NULL default value control phrase A X X X PRIMARY KEY table constraint attribute A X X X REFERENCES table constraint attribute A X X X TITLE output format phrase T X X X UC, UPPERCASE phrase T X X X Data Type / Literal ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions 280 SQL Reference: Fundamentals Functions, Operators, and Expressions The following list contains all SQL functions, operators, and expressions for this version and previous versions of Teradata Database. The following type codes appear in the ANSI Compliance column: Data Type Attributes, continued UNIQUE table constraint attribute A X X X WITH CHECK OPTION/ WITH NO CHECK OPTION column constraint attribute T X X X WITH DEFAULT default value control phrase T X X X Data Type / Literal ANSI Compliance V2R6.2 V2R6.1 V2R6.0 Code Definition A ANSI P Partially ANSI-compliant T Teradata extension Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0 - (subtract) A X X X - (unary minus) A X X X * (multiply) A X X X ** (exponentiate) T X X X / (divide) A X X X ^= (inequality) T X X X + (add) A X X X + (unary plus) A X X X < (less than) A X X X <= (less than or equal) A X X X <> (inequality) A X X XAppendix E: SQL Feature Summary Functions, Operators, and Expressions SQL Reference: Fundamentals 281 = (equality) A X X X > (greater than) A X X X >= (greater than or equal) A X X X ABS T X X X ACCOUNT T X X X ACOS T X X X ACOSH T X X X ADD_MONTHS T X X X ALL A X X X AND A X X X ANY A X X X ASIN T X X X ASINH T X X X ATAN T X X X ATAN2 T X X X ATANH T X X X AVE/ AVERAGE/ T X X X AVG A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X ROWS window_frame_extent A X X X BETWEEN NOT BETWEEN A X X X BYTE/ BYTES T X X X CASE A X X X CASE_N T X X X CAST A, T X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions 282 SQL Reference: Fundamentals CHAR/ CHARACTERS/ CHARS T X X X CHAR_LENGTH/ CHARACTER_LENGTH A X X X CHAR2HEXINT T X X X COALESCE A X X X CORR A X X X COS T X X X COSH T X X X COUNT A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X ROWS window_frame_extent A X X X COVAR_POP A X X X COVAR_SAMP A X X X CSUM T X X X CURRENT_DATE A X X X CURRENT_TIME A X X X CURRENT_TIMESTAMP A X X X DATABASE T X X X DATE T X X X DEFAULT A, T X EQ T X X X EXCEPT A, T X X X Options ALL T X X X EXISTS NOT EXISTS A X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions SQL Reference: Fundamentals 283 EXP T X X X EXTRACT P X X X FORMAT T X X X GE T X X X GROUPING A X X X GT T X X X HASHAMP T X X X HASHBAKAMP T X X X HASHBUCKET T X X X HASHROW T X X X IN NOT IN A X X X INDEX T X X X INTERSECT A, T X X X Options ALL T X X X IS NULL IS NOT NULL A X X X KURTOSIS A X X X LE T X X X LIKE NOT LIKE A X X X LN T X X X LOG T X X X LOWER A X X X LT T X X X MAVG T X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions 284 SQL Reference: Fundamentals MAX/ MAXIMUM A T X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X ROWS window_frame_extent A X X X MCHARACTERS T X X X MDIFF T X X X MIN/ MINIMUM A T X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X ROWS window_frame_extent A X X X MINUS T X X X Options ALL T X X X MLINREG T X X X MOD T X X X MSUM T X X X NE T X X X NEW P X X NOT A X X X NOT= T X X X NULLIF A X X X NULLIFZERO T X X X OCTET_LENGTH A X X X OR A X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions SQL Reference: Fundamentals 285 OVERLAPS A X X X PERCENT_RANK A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X POSITION A X X X PROFILE T X X X QUANTILE T X X X RANDOM T X X X RANGE_N T X X X RANK T X X X RANK A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X REGR_AVGX A X X X REGR_AVGY A X X X REGR_COUNT A X X X REGR_INTERCEPT A X X X REGR_R2 A X X X REGR_SLOPE A X X X REGR_SXX A X X X REGR_SXY A X X X REGR_SYY A X X X ROLE T X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions 286 SQL Reference: Fundamentals ROW_NUMBER A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X SESSION T X X X SIN T X X X SINH T X X X SKEW A X X X SOME A X X X SOUNDEX T X X X SQRT T X X X STDDEV_POP A X X X STDDEV_SAMP A X X X SUBSTR T X X X SUBSTRING A X X X SUM A X X X Options OVER A X X X PARTITION BY value_expression A X X X ORDER BY value_expression A X X X ROWS window_frame_extent A X X X TAN T X X X TANH T X X X TIME T X X X TITLE T X X X TRANSLATE A X X X TRANSLATE_CHK T X X X TRIM P X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions SQL Reference: Fundamentals 287 TYPE T X X X UNION A, T X X X Options ALL T X X X UPPER A X X X USER A X X X VAR_POP A X X X VAR_SAMP A X X X VARGRAPHIC T X X X WIDTH_BUCKET A X X X ZEROIFNULL T X X X Function / Operator / Expression ANSI Compliance V2R6.2 V2R6.1 V2R6.0Appendix E: SQL Feature Summary Functions, Operators, and Expressions 288 SQL Reference: FundamentalsSQL Reference: Fundamentals 289 Glossary AMP Access Module Processor vproc ANSI American National Standards Institute BLOB Binary Large Object BTEQ Basic TEradata Query facility BYNET Banyan Network - High speed interconnect CJK Chinese, Japanese, and Korean CLIv2 Call Level Interface Version 2 CLOB Character Large Object cs0, cs1, cs2, cs3 Four code sets (codeset 0, 1, 2, and 3) used in EUC encoding. distinct type A UDT that is based on a single predefined data type E2I External-to-Internal EUC Extended UNIX Code external routine UDF, UDM, or external stored procedure that is written using C or C++ external stored procedure a stored procedure that is written using C or C++ FK Foreign Key HI Hash Index I2E Internal-to-External JI Join Index JIS Japanese Industrial Standards LOB Large Object LT/ST Large Table/Small Table (join) NPPI Non-Partitioned Primary Index NUPI Non-Unique Primary Index NUSI Non-Unique Secondary Index OLAP On-Line Analytical Processing OLTP On-Line Transaction ProcessingGlossary 290 SQL Reference: Fundamentals QCD Query Capture Database PDE Parallel Database Extensions PE Parsing Engine vproc PI Primary Index PK Primary Key PPI Partitioned Primary Index predefined type Teradata Database system type such as INTEGER and VARCHAR RDBMS Relational Database Management System SDF Specification for Data Formatting stored procedure a stored procedure that is written using SQL statements structured type A UDT that is a collection of one or more fields called attributes, each of which is defined as a predefined data type or other UDT (which allows nesting) UCS-2 Universal Coded Character Set containing 2 bytes UDF User-Defined Function UDM User-Defined Method UDT User-Defined Type UPI Unique Primary Index USI Unique Secondary Index vproc Virtual ProcessSQL Reference: Fundamentals 291 Index Numerics 2PC, request processing 124 A ABORT statement 220 ABS function 281 ACCOUNT function 281 Account priority 141 ACOS function 281 ACOSH function 281 ACTIVITY_COUNT 144 ADD_MONTHS function 281 Aggregate join index 31 Aggregates, null and 137 ALL predicate 281 ALTER FUNCTION statement 220 ALTER METHOD statement 220 ALTER PROCEDURE statement 220 ALTER REPLICATION GROUP statement 221 ALTER SPECIFIC FUNCTION statement 220 ALTER SPECIFIC METHOD statement 220 ALTER TABLE statement 103, 221 ALTER TRIGGER statement 223 ALTER TYPE statement 223 Alternate key 37 AND operator 281 ANSI compliance and 218 ANSI DateTime, null and 134 ANSI SQL differences 218 Teradata compliance with 214 Teradata extensions to 218 Teradata terminology and 216 terminology differences 216 ANY predicate 281 ARC hash indexes and 35 join indexes and 32 referential integrity and 41 Archive and Recovery. See ARC Arithmetic function, nulls and 134 Arithmetic operators, nulls and 134 AS data type attribute 279 ASCII session character set 139 ASIN function 281 ASINH function 281 ASYNC statement modifier 275 ATAN function 281 ATAN2 function 281 ATANH function 281 AVE function 281 AVERAGE function 281 AVG function 281 B BEGIN DECLARE SECTION statement 223 BEGIN LOGGING statement 224 BEGIN QUERY LOGGING statement 224 BEGIN TRANSACTION statement 225 BETWEEN predicate 281 BIGINT data type 277 BINARY LARGE OBJECT. See BLOB BLOB data type 277 BYTE data type 277 Byte data types 15 BYTE function 281 BYTEINT data type 277 BYTES function 281 C CALL statement 225 Call-Level Interface. See CLI Cardinality, defined 2 CASE expression 281 CASE_N function 281 CASESPECIFIC data type attribute 279 CAST function 281 CD-ROM images v CHAR data type 277 CHAR function 282 CHAR VARYING data type 277 CHAR_LENGTH function 282 CHAR2HEXINT function 282 Character data literal 278 CHARACTER data type 277 Character data types 13 CHARACTER LARGE OBJECT. See CLOB Character literals 89 Character names 77 CHARACTER SET data type attribute 279 Character set, request change of 142 Character sets, Teradata SQL lexicon 67Index 292 SQL Reference: Fundamentals CHARACTER VARYING data type 277 CHARACTER_LENGTH function 282 CHARACTERS function 282 CHARS function 282 CHECK data type attribute 279 CHECKPOINT statement 225 Child table, defined 37 Circular reference, referential integrity 39 Classes of UDFs aggregate 54 scalar 54 CLI session management 143 CLOB data type 277 CLOSE statement 225 COALESCE expression 282 Collation sequences (SQL) 140 COLLECT DEMOGRAPHICS statement 225 COLLECT STAT INDEX statement 227 COLLECT STAT statement 226 COLLECT STATISTICS INDEX statement 227 COLLECT STATISTICS statement 226 COLLECT STATS INDEX statement 227 COLLECT STATS statement 226 Collecting statistics 164 Column alias 72 Columns definition 12 referencing, syntax for 72 COMMENT statement 227 Comments bracketed 96 multibyte character sets and 96 simple 95 COMMIT statement 228 Comparison operators, null and 135 COMPRESS data type attribute 279 CONNECT statement 228 Constants. See Literals CONSTRAINT data type attribute 279 CORR function 282 COS function 282 COSH function 282 COVAR_SAMP function 282 Covering index 31 Covering, secondary index, non-unique, and 27 CREATE AUTHORIZATION statement 228 CREATE CAST statement 229 CREATE DATABASE statement 229 CREATE FUNCTION statement 55, 229 CREATE HASH INDEX statement 231 CREATE INDEX statement 232 CREATE JOIN INDEX statement 232 CREATE MACRO statement 233 CREATE METHOD statement 233 CREATE ORDERING statement 234 CREATE PROCEDURE statement 53, 234 CREATE PROFILE statement 236 CREATE RECURSIVE VIEW statement 244 CREATE REPLICATION GROUP statement 237 CREATE ROLE statement 237 CREATE TABLE statement 237 CREATE TRANSFORM statement 239 CREATE TRIGGER statement 239 CREATE TYPE statement 240, 241 CREATE USER statement 242 CREATE VIEW statement 243 CS data type attribute 279 CSUM function 282 CURRENT_DATE function 282 CURRENT_TIME function 282 CURRENT_TIMESTAMP function 282 Cylinder reads 164 D Data Control Language. See DCL Data Definition Language. See DDL Data Manipulation Language. See DML Data types byte 15 character 13 DateTime 14 definition 13 interval 14 numeric 13 UDT 15, 58 Data, standard form of, Teradata Database 71 Database default, establishing for session 76 default, establishing permanent 75 DATABASE function 282 DATABASE statement 244 Database, defined 1 DATE data type 277 DATE function 282 DATE literal 278 Date literals 88 Date, change format of 142 DateTime data types 14 DCL statements, defined 105 DDL CREATE FUNCTION 55 CREATE PROCEDURE 53 REPLACE FUNCTION 55 REPLACE PROCEDURE 53 DDL statements, defined 101 DEC data type 277Index SQL Reference: Fundamentals 293 DECIMAL data type 277 Decimal literal 278 Decimal literals 87 DECLARE CURSOR statement 244 DECLARE STATEMENT statement 244 DECLARE TABLE statement 244 DEFAULT data type attribute 279 DEFAULT function 282 Degree, defined 2 DELETE DATABASE statement 245 DELETE statement 245 DELETE USER statement 245 Delimiters 93 DESCRIBE statement 246 DIAGNOSTIC "validate index" statement 246 DIAGNOSTIC DUMP SAMPLES statement 246 DIAGNOSTIC HELP SAMPLES statement 246 DIAGNOSTIC SET SAMPLES statement 246 Distinct UDTs 58 DML statements, defined 106 DOUBLE PRECISION data type 277 DROP AUTHORIZATION statement 246 DROP CAST statement 246 DROP DATABASE statement 246 DROP FUNCTION statement 246 DROP HASH INDEX statement 246 DROP INDEX statement 247 DROP JOIN INDEX statement 247 DROP MACRO statement 247 DROP ORDERING statement 247 DROP PROCEDURE statement 247 DROP PROFILE statement 247, 256 DROP REPLICATION GROUP statement 247 DROP ROLE statement 247 DROP SPECIFIC FUNCTION statement 246 DROP STATISTICS statement 247 DROP TABLE statement 248 DROP TRANSFORM statement 248 DROP TRIGGER statement 248 DROP TYPE statement 248 DROP USER statement 246 DROP VIEW statement 248 DUMP EXPLAIN statement 248 E EBCDIC session character set 139 ECHO statement 248 Embedded SQL binding style 100 macros 46 END DECLARE SECTION statement 248 END LOGGING statement 249 END QUERY LOGGING statement 249 END TRANSACTION statement 249 END-EXEC statement 248 EQ operator 282 Event processing SELECT AND CONSUME and 133 EXCEPT operator 282 EXEC SQL statement modifier 275 Executable SQL statements 119 EXECUTE IMMEDIATE statement 249 EXECUTE statement 249 EXISTS predicate 282 EXP function 283 EXPLAIN request modifier 19, 21, 275 Express logon 142 External stored procedures 53 usage 53 EXTRACT function 283 F Fallback hash indexes and 35 join indexes and 32 FastLoad hash indexes and 35 join indexes and 32 referential integrity and 42 FETCH statement 250 FLOAT data type 277 Floating point literal 278 Floating point literals 87 Foreign key defined 16 maintaining 40 FOREIGN KEY data type attribute 279 Foreign key. See also Key Foreign key. See also Referential integrity FORMAT data type attribute 279 FORMAT function 283 Full table scan 163 G GE operator 283 general information about Teradata vi GET CRASH statement 250 GIVE statement 250 GRANT statement 250 GRAPHIC data type 277 Graphic literal 278 Graphic literals 89 GROUP BY statement modifier 275 GROUPING function 283 GT operator 283Index 294 SQL Reference: Fundamentals H Hash buckets 18 Hash index ARC and 35 effects of 35 MultiLoad and 35 permanent journal and 35 TPump and 35 Hash mapping 18 HASHAMP function 283 HASHBAKAMP function 283 HASHBUCKET function 283 HASHROW function 283 HAVING statement modifier 275 HELP statement 252 HELP statements 116 HELP STATISTICS statement 253 Hexadecimal get representation of name 84 Hexadecimal literal 278 Hexadecimal literals 87 I IN predicate 283 INCLUDE SQLCA statement 254 INCLUDE SQLDA statement 254 INCLUDE statement 254 Index advantages of 18 covering 31 defined 17 disadvantages of 18 dropping 105 EXPLAIN, using 21 hash mapping and 18 join 20 keys and 16 maximum number of columns 206 non-unique 19 partitioned 20 row hash value and 17 RowID and 17 selectivity of 17 types of (Teradata) 19 unique 19 uniqueness value and 17 INDEX function 283 Information Products Publishing Library v INITIATE INDEX ANALYSIS statement 254 INSERT EXPLAIN statement 255 INSERT statement 254 INT data type 277 INTEGER data type 277 Integer literal 278 Integer literals 87 INTERSECT operator 283 Interval data types 14 INTERVAL DAY data type 277 INTERVAL DAY TO HOUR data type 277 INTERVAL DAY TO MINUTE data type 277 INTERVAL DAY TO SECOND data type 277 INTERVAL HOUR data type 277 INTERVAL HOUR TO MINUTE data type 277 INTERVAL HOUR TO SECOND data type 278 Interval literal 279 Interval literals 88 INTERVAL MINUTE data type 278 INTERVAL MINUTE TO SECOND data type 278 INTERVAL MONTH data type 278 INTERVAL SECOND data type 278 INTERVAL YEAR data type 278 INTERVAL YEAR TO MONTH data type 278 IS NOT NULL predicate 283 IS NULL predicate 283 Iterated requests 127 J Japanese character code notation, how to read 171 Japanese character names 77 JDBC 100 Join index aggregate 31 described 30 effects of 32 multitable 31 performance and 33 queries using 33 single-table 31 sparse 32 Join Index. See also Index K Key alternate 37 foreign 16 indexes and 16 primary 16 referential integrity and 16 Keywords 66 NULL 90 KURTOSIS function 283 L LE operator 283 Lexical separators 94Index SQL Reference: Fundamentals 295 LIKE predicate 283 Limits database 206 session 211 system 204 Literals character 89 date 88 decimal 87 floating point 87 graphic 89 hexadecimal 87 integer 87 interval 88 time 88 timestamp 88 LN function 283 LOCKING statement modifier 275 LOG function 283 LOGOFF statement 255 LOGON statement 255 Logon, express 142 LONG VARCHAR data type 278 LONG VARGRAPHIC data type 278 LOWER function 283 LT operator 283 M Macros contents 47 defined 46 executing 47 maximum expanded text size 207 maximum number of parameters 207 SQL statements and 46 MAVG function 283 MAX function 284 MAXIMUM function 284 MCHARACTERS function 284 MDIFF function 284 MERGE statement 255 MIN function 284 MINIMUM function 284 MINUS operator 284 MLINREG function 284 MOD operator 284 MODIFY DATABASE statement 256 MODIFY USER statement 257 MSUM function 284 MultiLoad hash indexes and 35 join indexes and 32 referential integrity and 42 Multi-statement requests, performance 125 Multi-statement transactions 125 Multitable join index 31 N Name calculate length of 78 fully qualified 72 get hexadecimal representation 84 identify in logon string 86 maximum size 206 multiword 69 object 77 resolving 74 translation and storage 81 NAMED data type attribute 279 NE operator 284 NEW expression 284 Nonexecutable SQL statements 120 Non-partitioned primary index. See NPPI. Non-unique index. See Index, Primary index, Secondary index NOT BETWEEN predicate 281 NOT CASESPECIFIC data type attribute 279 NOT CS data type attribute 279 NOT EXISTS predicate 282 NOT IN predicate 283 NOT LIKE predicate 283 NOT NULL data type attribute 279 NOT operator 284 NOT= operator 284 NPPI 20 Null aggregates and 137 ANSI DateTime and 134 arithmetic functions and 134 arithmetic operators and 134 collation sequence 136 comparison operators and 135 excluding 135 operations on (SQL) 134 searching for 136 searching for, null and non-null 136 NULL keyword 90 Null statement 98 NULLIF expression 284 NULLIFZERO function 284 NUMERIC data type 278 Numeric data types 13 NUPI. See Primary index, non-unique NUSI. See Secondary index, non-uniqueIndex 296 SQL Reference: Fundamentals O Object names 77 Object, name comparison 82 OCTET_LENGTH function 284 ODBC 100 OPEN statement 258 Operators 91 OR operator 284 ORDER BY statement modifier 276 ordering publications v OVERLAPS operator 285 P Parallel step processing 125 Parameters, session 138 Parent table, defined 37 Partial cover 30 Partition elimination 159 Partitioned primary index. See PPI. PERCENT_RANK function 285 Permanent journal creating 2 hash indexes and 35 join indexes and 32 POSITION function 285 POSITION statement 258 PPI defined 20 maximum number of partitions 206 partition elimination and 159 Precedence, SQL operators 91 PREPARE statement 258 Primary index choosing 23 default 22 described 22 non-unique 23 NULL and 136 summary 24 unique 23 PRIMARY KEY data type attribute 279 Primary key, defined 16 Primary key. See also Key Procedure, dropping 105 product-related information v PROFILE function 285 Profiles 55 publications related to this release v Q QCD tables populating 115 QUALIFY statement modifier 276 QUANTILE function 285 Query Capture Database. See QCD Query processing access types 162 all AMP request 156 AMP sort 158 BYNET merge 159 defined 153 full table scan 163 single AMP request 154 single AMP response 156 Query, defined 153 R RANDOM function 285 RANGE_N function 285 RANK function 285 REAL data type 278 Recursive queries (SQL) 112 Recursive query, defined 112 REFERENCES data type attribute 279 Referential integrity ARC and 41 circular references and 39 described 36 FastLoad and 42 foreign keys and 39 importance of 38 MultiLoad and 42 terminology 37 REGR_AVGX function 285 REGR_AVGY function 285 REGR_COUNT function 285 REGR_INTERCEPT function 285 REGR_R2 function 285 REGR_SLOPE function 285 REGR_SXX function 285 REGR_SXY function 285 REGR_SYY function 285 release definition v RENAME FUNCTION statement 259 RENAME MACRO statement 259 RENAME PROCEDURE statement 259 RENAME TABLE statement 259 RENAME TRIGGER statement 259 RENAME VIEW statement 259 REPLACE CAST statement 259 REPLACE FUNCTION statement 55, 259 REPLACE MACRO statement 261 REPLACE METHOD statement 261 REPLACE ORDERING statement 262 REPLACE PROCEDURE statement 53, 262, 263Index SQL Reference: Fundamentals 297 REPLACE TRANSFORM statement 265 REPLACE TRIGGER statement 266 REPLACE VIEW statement 266 Request processing 2PC 124 ANSI mode 123 Teradata mode 123 Request terminator 96 Requests iterated 127 maximum size 207 multi-statement 120 single-statement 120 Requests. See also Blocked requests, Multi-statement requests, Request processing Reserved words 219 RESTART INDEX ANALYSIS statement 266 Restricted words 173 REVOKE statement 266 REWIND statement 268 ROLE function 285 Roles 57 ROLLBACK statement 268 ROW_NUMBER function 286 Rows, maximum size 206 S SAMPLE statement modifier 276 Secondary index defined 25 dual 28 non-unique 26 bit mapping 28 covering and 27 value-ordered 27 NULL and 136 summary 29 unique 26 using Teradata Index Wizard 21 Security, user-level password attributes 56 Seed statements 113 SELECT statement 268 Selectivity high 17 low 17 Semicolon null statement 98 request terminator 96 statement separator 94 Separator lexical 94 statement 94 Session character set ASCII 139 EBCDIC 139 UTF16 139 UTF8 139 Session collation 140 Session control 138 SESSION function 286 Session handling, session control 144 Session management CLI 143 ODBC 143 requests 144 session reserve 143 Session parameters 138 SET BUFFERSIZE statement 270 SET CHARSET statement 270 SET CONNECTION statement 270 SET CRASH statement 270 SET ROLE statement 271 SET SESSION ACCOUNT statement 271 SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL statement 271 SET SESSION COLLATION statement 271 SET SESSION DATABASE statement 271 SET SESSION DATEFORM statement 271 SET SESSION FUNCTION TRACE statement 271 SET SESSION OVERRIDE REPLICATION statement 272 SET SESSION statement 271 SET TIME ZONE statement 272 SHOW CAST statement 272 SHOW FUNCTION statement 272 SHOW HASH INDEX statement 272 SHOW JOIN INDEX statement 272 SHOW MACRO statement 272 SHOW METHOD statement 272 SHOW PROCEDURE statement 272 SHOW REPLICATION GROUP statement 272 SHOW SPECIFIC FUNCTION statement 272 SHOW statement 272 SHOW statements 117 SHOW TABLE statement 272 SHOW TRIGGER statement 272 SHOW TYPE statement 272 SHOW VIEW statement 272 SIN function 286 Single-table join index 31 SINH function 286 SKEW function 286 SMALLINT data type 278 SOME predicate 286 SOUNDEX function 286 Sparse join index 32Index 298 SQL Reference: Fundamentals Specifications database 206 session 211 system 204 SQL dynamic 129 dynamic, SELECT statement and 131 static 129 SQL binding styles CLI 100 defined 100 direct 100 embedded 100 JDBC 100 ODBC 100 stored procedure 100 SQL data type attributes AS 279 CASESPECIFIC 279 CHARACTER SET 279 CHECK 279 COMPRESS 279 CONSTRAINT 279 CS 279 DEFAULT 279 FOREIGN KEY 279 FORMAT 279 NAMED 279 NOT CASESPECIFIC 279 NOT CS 279 NOT NULL 279 PRIMARY KEY 279 REFERENCES 279 TITLE 279 UC 279 UNIQUE 280 UPPERCASE 279 WITH CHECK OPTION 280 WITH DEFAULT 280 SQL data types BIGINT 277 BLOB 277 BYTE 277 BYTEINT 277 CHAR 277 CHAR VARYING 277 CHARACTER 277 CHARACTER VARYING 277 CLOB 277 DATE 277 DEC 277 DECIMAL 277 DOUBLE PRECISION 277 FLOAT 277 GRAPHIC 277 INT 277 INTEGER 277 INTERVAL DAY 277 INTERVAL DAY TO HOUR 277 INTERVAL DAY TO MINUTE 277 INTERVAL DAY TO SECOND 277 INTERVAL HOUR 277 INTERVAL HOUR TO MINUTE 277 INTERVAL HOUR TO SECOND 278 INTERVAL MINUTE 278 INTERVAL MINUTE TO SECOND 278 INTERVAL MONTH 278 INTERVAL SECOND 278 INTERVAL YEAR 278 INTERVAL YEAR TO MONTH 278 LONG VARCHAR 278 LONG VARGRAPHIC 278 NUMERIC 278 REAL 278 SMALLINT 278 TIME 278 TIME WITH TIMEZONE 278 TIMESTAMP 278 TIMESTAMP WITH TIMEZONE 278 UDT 278 VARBYTE 278 VARCHAR 278 VARGRAPHIC 278 SQL error response (ANSI) 149 SQL expressions CASE 281 COALESCE 282 NEW 284 NULLIF 284 SQL Flagger enabling and disabling 217 function 217 session control 139 SQL functional families, defined 99 SQL functions ABS 281 ACCOUNT 281 ACOS 281 ACOSH 281 ADD_MONTHS 281 ASIN 281 ASINH 281 ATAN 281 ATAN2 281 ATANH 281 AVE 281 AVERAGE 281 AVG 281Index SQL Reference: Fundamentals 299 BYTE 281 BYTES 281 CASE_N 281 CAST 281 CHAR 282 CHAR_LENGTH 282 CHAR2HEXINT 282 CHARACTER_LENGTH 282 CHARACTERS 282 CHARS 282 CORR 282 COS 282 COSH 282 COVAR_SAMP 282 CSUM 282 CURRENT_DATE 282 CURRENT_TIME 282 CURRENT_TIMESTAMP 282 DATABASE 282 DATE 282 DEFAULT 282 EXP 283 EXTRACT 283 FORMAT 283 GROUPING 283 HASHAMP 283 HASHBAKAMP 283 HASHBUCKET 283 HASHROW 283 INDEX 283 KURTOSIS 283 LN 283 LOG 283 LOWER 283 MAVG 283 MAX 284 MAXIMUM 284 MCHARACTERS 284 MDIFF 284 MIN 284 MINIMUM 284 MLINREG 284 MSUM 284 NULLIFZERO 284 OCTET_LENGTH 284 PERCENT_RANK 285 POSITION 285 PROFILE 285 QUANTILE 285 RANDOM 285 RANGE_N 285 RANK 285 REGR_AVGX 285 REGR_AVGY 285 REGR_COUNT 285 REGR_INTERCEPT 285 REGR_R2 285 REGR_SLOPE 285 REGR_SXX 285 REGR_SXY 285 REGR_SYY 285 ROLE 285 ROW_NUMBER 286 SESSION 286 SIN 286 SINH 286 SKEW 286 SOUNDEX 286 SQRT 286 STDDEV_POP 286 STDDEV_SAMP 286 SUBSTR 286 SUBSTRING 286 SUM 286 TAN 286 TANH 286 TIME 286 TITLE 286 TRANSLATE 286 TRANSLATE_CHK 286 TRIM 286 TYPE 287 UNION 287 UPPER 287 USER 287 VAR_POP 287 VAR_SAMP 287 VARGRAPHIC 287 WIDTH_BUCKET 287 ZEROIFNULL 287 SQL lexicon character names 77 delimiters 93 Japanese character names 67, 77 keywords 66 lexical separators 94 object names 77 operators 91 request terminator 96 statement separator 94 SQL literals Character data 278 DATE 278 Decimal 278 Floating point 278 Graphic 278 Hexadecimal 278 Integer 278Index 300 SQL Reference: Fundamentals Interval 279 TIME 279 TIMESTAMP 279 SQL operators AND 281 EQ 282 EXCEPT 282 GE 283 GT 283 INTERSECT 283 LE 283 LT 283 MINUS 284 MOD 284 NE 284 NOT 284 NOT= 284 OR 284 OVERLAPS 285 SQL predicates ALL 281 ANY 281 BETWEEN 281 EXISTS 282 IN 283 IS NOT NULL 283 IS NULL 283 LIKE 283 NOT BETWEEN 281 NOT EXISTS 282 NOT IN 283 NOT LIKE 283 SOME 286 SQL request modifier, EXPLAIN 19, 21, 275 SQL requests iterated 127 multi-statement 120 single-statement 120 SQL responses 147 failure 150 success 148 warning 149 SQL return codes 144 SQL statement modifiers ASYNC 275 EXEC SQL 275 GROUP BY 275 HAVING 275 LOCKING 275 ORDER BY 276 QUALIFY 276 SAMPLE 276 USING 276 WHERE 276 WITH 276 WITH RECURSIVE 276 SQL statements ABORT 220 ALTER FUNCTION 220 ALTER METHOD 220 ALTER PROCEDURE 220 ALTER REPLICATION GROUP 221 ALTER SPECIFIC FUNCTION 220 ALTER SPECIFIC METHOD 220 ALTER TABLE 221 ALTER TRIGGER 223 ALTER TYPE 223 BEGIN DECLARE SECTION 223 BEGIN LOGGING 224 BEGIN QUERY LOGGING 224 BEGIN TRANSACTION 225 CALL 225 CHECKPOINT 225 CLOSE 225 COLLECT DEMOGRAPHICS 225 COLLECT STAT 226 COLLECT STAT INDEX 227 COLLECT STATISTICS 226 COLLECT STATISTICS INDEX 227 COLLECT STATS 226 COLLECT STATS INDEX 227 COMMENT 227 COMMIT 228 CONNECT 228 CREATE AUTHORIZATION 228 CREATE CAST 229 CREATE DATABASE 229 CREATE FUNCTION 229 CREATE HASH INDEX 231 CREATE INDEX 232 CREATE JOIN INDEX 232 CREATE MACRO 233 CREATE METHOD 233 CREATE ORDERING 234 CREATE PROCEDURE 234 CREATE PROFILE 236 CREATE RECURSIVE VIEW 244 CREATE REPLICATION GROUP 237 CREATE ROLE 237 CREATE TABLE 237 CREATE TRANSFORM 239 CREATE TRIGGER 239 CREATE TYPE 240, 241 CREATE USER 242 CREATE VIEW 243 DATABASE 244 DECLARE CURSOR 244 DECLARE STATEMENT 244Index SQL Reference: Fundamentals 301 DECLARE TABLE 244 DELETE 245 DELETE DATABASE 245 DELETE USER 245 DESCRIBE 246 DIAGNOSTIC 115, 246 DIAGNOSTIC "validate index" 246 DIAGNOSTIC DUMP SAMPLES 246 DIAGNOSTIC HELP SAMPLES 246 DIAGNOSTIC SET SAMPLES 246 DROP AUTHORIZATION 246 DROP CAST 246 DROP DATABASE 246 DROP FUNCTION 246 DROP HASH INDEX 246 DROP INDEX 247 DROP JOIN INDEX 247 DROP MACRO 247 DROP ORDERING 247 DROP PROCEDURE 247 DROP PROFILE 247, 256 DROP REPLICATION GROUP 247 DROP ROLE 247 DROP SPECIFIC FUNCTION 246 DROP STATISTICS 247 DROP TABLE 248 DROP TRANSFORM 248 DROP TRIGGER 248 DROP TYPE 248 DROP USER 246 DROP VIEW 248 DUMP EXPLAIN 248 ECHO 248 END DECLARE SECTION 248 END LOGGING 249 END QUERY LOGGING 249 END TRANSACTION 249 END-EXEC 248 executable 119 EXECUTE 249 EXECUTE IMMEDIATE 249 FETCH 250 GET CRASH 250 GIVE 250 GRANT 250 HELP 252 HELP STATISTICS 253 INCLUDE 254 INCLUDE SQLCA 254 INCLUDE SQLDA 254 INITIATE INDEX ANALYSIS 254 INSERT 254 INSERT EXPLAIN 255 invoking 119 LOGOFF 255 LOGON 255 MERGE 255 MODIFY DATABASE 256 MODIFY USER 257 name resolution 74 nonexecutable 120 OPEN 258 partial names, use of 73 POSITION 258 PREPARE 258 RENAME FUNCTION 259 RENAME MACRO 259 RENAME PROCEDURE 259 RENAME TABLE 259 RENAME TRIGGER 259 RENAME VIEW 259 REPLACE CAST 259 REPLACE FUNCTION 259 REPLACE MACRO 261 REPLACE METHOD 261 REPLACE ORDERING 262 REPLACE PROCEDURE 262, 263 REPLACE TRANSFORM 265 REPLACE TRIGGER 266 REPLACE VIEW 266 RESTART INDEX ANALYSIS 266 REVOKE 266 REWIND 268 ROLLBACK 268 SELECT 268 SELECT, dynamic SQL 131 SET BUFFERSIZE 270 SET CHARSET 270 SET CONNECTION 270 SET CRASH 270 SET ROLE 271 SET SESSION 271 SET SESSION ACCOUNT 271 SET SESSION CHARACTERISTICS AS TRANSACTION ISOLATION LEVEL 271 SET SESSION COLLATION 271 SET SESSION DATABASE 271 SET SESSION DATEFORM 271 SET SESSION FUNCTION TRACE 271 SET SESSION OVERRIDE REPLICATION 272 SET TIME ZONE 272 SHOW 272 SHOW CAST 272 SHOW FUNCTION 272 SHOW HASH INDEX 272 SHOW JOIN INDEX 272 SHOW MACRO 272 SHOW METHOD 272Index 302 SQL Reference: Fundamentals SHOW PROCEDURE 272 SHOW REPLICATION GROUP 272 SHOW SPECIFIC FUNCTION 272 SHOW TABLE 272 SHOW TRIGGER 272 SHOW TYPE 272 SHOW VIEW 272 structure 63 subqueries 110 TEST 273 UPDATE 273 WAIT 274 WHENEVER 274 SQL statements, macros and 46 SQL. See also Embedded SQL SQL-2003 non-reserved words 174 SQL-2003 reserved words 174 SQLCA 144 SQLCODE 144 SQLSTATE 144 SQRT function 286 Statement processing. See Query processing Statement separator 94 STDDEV_POP function 286 STDDEV_SAMP function 286 Stored procedures ACTIVITY_COUNT 144 creating 50 deleting 52 elements of 49 executing 51 modifying 51 privileges 49 renaming 52 Structured UDTs 58 Subqueries (SQL) 110 Subquery, defined 110 SUBSTR function 286 SUBSTRING function 286 SUM function 286 Syntax, how to read 167 T Table cardinality of 2 creating indexes for 20 defined 2 degree of 2 dropping 105 full table scan 163 global temporary 5 global temporary trace 4 maximum number of columns 206 maximum number of rows 206 queue 4 tuple and 2 volatile temporary 9 Table structure, altering 103 Table, change structure of 103 TAN function 286 TANH function 286 Target level emulation 115 Teradata Database database specifications 206 session specifications 211 system specifications 204 Teradata DBS, session management 143 Teradata Index Wizard 21 determining optimum secondary indexes 21 SQL diagnostic statements 115 Teradata SQL 218 Teradata SQL, ANSI SQL and 214 Terminator, request 96 TEST statement 273 TIME data type 278 TIME function 286 TIME literal 279 Time literals 88 TIME WITH TIMEZONE data type 278 TIMESTAMP data type 278 TIMESTAMP literal 279 Timestamp literals 88 TIMESTAMP WITH TIMEZONE data type 278 TITLE data type attribute 279 TITLE function 286 TITLE phrase, column definition 71 TPump hash indexes and 35 join indexes and 32 Transaction mode, session control 140 Transaction modes (SQL) 140 Transactions defined 122 explicit, defined 124 implicit, defined 124 TRANSLATE function 286 TRANSLATE_CHK function 286 Trigger altering 44 creating 44 defined 44 dropping 44, 105 process flow for 44 TRIM function 286 Two-phase commit. See 2PC TYPE function 287Index SQL Reference: Fundamentals 303 U UC data type attribute 279 UDFs classes 54 CREATE FUNCTION 55 CREATE PROCEDURE 53 usage 55 UDT data types 15, 58, 278 creating and using 59 distinct 58 structured 58 Unicode, notation 171 UNION function 287 UNIQUE alternate key 37 UNIQUE data type attribute 280 Unique index. See Index, Primary index, Secondary index UPDATE statement 273 UPI. See Primary index, unique UPPER function 287 UPPERCASE data type attribute 279 USER function 287 User, defined 1 User-defined types. See UDT data types USI. See Secondary index, unique USING statement modifier 276 UTF16 session character set 139 UTF8 session character set 139 V VAR_POP function 287 VAR_SAMP function 287 VARBYTE data type 278 VARCHAR data type 278 VARGRAPHIC data type 278 VARGRAPHIC function 287 View described 42 dropping 105 maximum expanded text size 207 maximum number of columns 206 restrictions 43 W WAIT statement 274 WHENEVER statement 274 WHERE statement modifier 276 WIDTH_BUCKET function 287 WITH DEFAULT data type attribute 280 WITH NO CHECK OPTION data type attribute 280 WITH RECURSIVE statement modifier 276 WITH statement modifier 276 Z ZEROIFNULL function 287 Zero-table SELECT statement 108Index 304 SQL Reference: Fundamentals