Information Integration Solutions Center of Excellence

Parallel Framework Standard Practices
Investigate, Design, Develop: Data Flow Job Development
Prepared by IBM Information Integration Solutions Center of Excellence July 17, 2006

CONFIDENTIAL, PROPRIETARY, AND TRADE SECRET NATURE OF ATTACHED DOCUMENTS
This document is Confidential, Proprietary and Trade Secret Information (“Confidential Information”) of IBM, Inc. and is provided solely for the purpose of evaluating IBM products with the understanding that such Confidential Information will be disclosed only to those who have a “need to know.” The attached documents constitute Confidential Information as they include information relating to the business and/or products of IBM (including, without limitation, trade secrets, technical, business, and financial information) and are trade secret under the laws of the State of Massachusetts and the United States. Copyrights © 2006 IBM Information Integration Solutions All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM. While every precaution has been taken in the preparation of this document to reflect current information, IBM assumes no responsibility for errors or omissions or for damages resulting from the use of information contained herein.

Information Integration Solutions Center of Excellence

Document Goals
Intended Use This document presents a set of standard practices, methodologies, and examples for IBM WebSphere® DataStage Enterprise Edition™ (“DS/EE”) on UNIX, Windows, and USS. Except where noted, this document is intended to supplement, not replace the installation documentation. The primary audience for this document is DataStage developers who have been trained in Enterprise Edition. Information in certain sections may also be relevant for Technical Architects, System Administrators, and Developers This document is intended for the following product releases: - WebSphere DataStage Enterprise Edition 7.5.1 (UNIX, USS) - WebSphere DataStage Enterprise Edition 7.5x2 (Windows)

Target Audience Product Version

Document Revision History
Date
April 16, 2004 June 30, 2005 December 9, 2005 January 31, 2006 February 17, 2006 March 10, 2006 March 31, 2006

Rev.
1.0 2.0 3.0 3.1 4.0 4.1 4.2

Description
Initial Services release First version based on separation of EE BP into four separate documents, merged new material on Remote DB2, configuring DS for multiple users. Significant updates, additional material Updates based on review feedback. Added patch install checklist item (7.10) and Windows 7.5x2 patch list. Significant updates, new material on ETL overview, data types, naming standards, USS, design standards, database stage usage, database data type mappings, updated styles and use of cross-references. Corrected missing Figure 9. Added new material on establishing job boundaries, balancing job resource requirements / startup time with required data volume and processing windows, and minimizing number of runtime processes. Moved Baselining Performance discussion to Performance Tuniing BP. Expanded performance tuning section. Removed Architecture Overview (now a separate document). Expanded file stage recommendations. Updated directory naming standards for consistency with DS/EE Automation Standards and Toolkit. Segmented content into “Red Book” and “Standards”. Clarified terminology (“Best Practices”). Incorporated additional field feedback.

May 08, 2006 July 17, 2006

4.3 5.0

Document Conventions
This document uses the following conventions: Convention Usage Bold In syntax, bold indicates commands, function names, keywords, and options that must be input exactly as shown. In text, bold indicates keys to press, function names, and menu selections. Italic In syntax, italic indicates information that you supply. In text, italic also indicates UNIX commands and options, file names, and pathnames. Plain In text, plain indicates Windows NT commands and options, file names, and pathnames. Bold Italic Indicates: important information.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 2 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence Lucida Console Lucida Bold

Lucida Console text indicates examples of source code and system output. In examples, Lucida Console bold indicates characters that the user types or keys the user presses (for example, <Return>). In examples, Lucida Blue will be used to illustrate operating system command line prompt. A right arrow between menu commands indicates you should choose each command in sequence. For example, “Choose File Exit” means you should choose File from the menu bar, and then choose Exit from the File pull-down menu. The continuation character  is used in source code examples to indicate a line that is too long to fit on the page, but must be entered as a single line on screen.

Lucida Blue

This line  continues

The following are also used: • Syntax definitions and examples are indented for ease in reading. • All punctuation marks included in the syntax—for example, commas, parentheses, or quotation marks—are required unless otherwise indicated. • Syntax lines that do not fit on one line in this manual are continued on subsequent lines. The continuation lines are indented. When entering syntax, type the entire syntax entry, including the continuation lines, on the same input line. • Text enclosed in parenthesis and underlined (like this) following the first use of proper terms will be used instead of the proper term. Interaction with our example system will usually include the system prompt (in blue) and the command, most often on 2 or more lines. If appropriate, the system prompt will include the user name and directory for context. For example:
%etl_node%:dsadm /usr/dsadm/Ascential/DataStage > /bin/tar –cvf /dev/rmt0 /usr/dsadm/Ascential/DataStage/Projects

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

3 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

.......8 TUNING SORT.....................................................................................................................................................68 5..................................................................................................................3 LINK SORT AND SORT STAGE..................................59 5...................................................... ..............41 3...........................................................................................................................................................4 WORKING WITH SOURCE CODE CONTROL SYSTEMS..................3 JOB DESIGN TEMPLATES.............................................................................................................................................................................2 COMPLETE (TOTAL) SORT.................................73 6 SORTING.........................................................................................4 DEFAULT JOB DESIGN.....5 SUB-SORTS.........................................................................................................................................................................42 3......18 2......................................................2 ESTABLISHING JOB BOUNDARIES...............8 COMPONENT USAGE........................................................................................... 2006 4 of 179 © 2006 IBM Information Integration Solutions......................79 6................................................................................5 COLLECTOR TYPES.......................................................................................................................39 3..................................................................................................2 MONITORING PARTITIONS.......................77 6. stored in a retrieval system......................................................................................................3 DOCUMENTATION AND ANNOTATION.........29 2.........................................6 AUTOMATICALLY-INSERTED SORTS.....................72 5.........................................................................................5 JOB PARAMETERS..................43 3.......................................................................31 2........3 PARTITION METHODOLOGY...................................................................3 SEQUENTIAL FILE STAGES (IMPORT AND EXPORT).........2 JOB TYPES.....................................35 3 DEVELOPMENT GUIDELINES.............................................................................................6 PARALLEL SHARED CONTAINERS......................58 5 PARTITIONING AND COLLECTING............................6 COLLECTING METHODOLOGY..................................................................................................................................................................................51 4 DATASTAGE DATA TYPES...............4 STABLE SORT.............87 Parallel Framework Red Book: Data Flow Job Design July 17.................................................................................81 7.............5 UNDERSTANDING A JOB’S ENVIRONMENT...........................Information Integration Solutions Center of Excellence Table of Contents 1 DATA INTEGRATION OVERVIEW...................................................................................................................................................................................59 5...............70 5...................................................................................................................................67 5....................................................6 1.13 2........................................................................40 3.................................................................................................................................................................43 3.............................................................................................................................................85 8 TRANSFORMATION LANGUAGES............................. No part of this publication may be reproduced............82 7..................................................................................................................................................................................................................................................................................2 DATA SET USAGE.....................78 6....... or translated into any language in any form by any means without the written permission of IBM....................................................7 ERROR AND REJECT RECORD HANDLING..................................................39 3............................................. transmitted...............3 RUNTIME COLUMN PROPAGATION......................................................................................................................................................................................................................................................................................................... All rights reserved.....................................................1 WHICH FILE STAGE TO USE.............................................2 NULL HANDLING.......................................................................................................54 4...........................................................................................56 4...............................................................................................................................................................................................................................................................................................76 6...........................................................................................7 SORT METHODOLOGY.............................................................................1 PARTITION AND SORT KEYS............................................................................. transcribed.........74 6.......................................................................................................................2 NAMING CONVENTIONS..........................................................................................................................................................79 7 FILE STAGE USAGE........................................................1 JOB SEQUENCES......................................................74 6...................13 2.............................................39 3.......................................................................4 PARTITIONING EXAMPLES................................81 7.......................7 1.........................75 6..........................................................................................................................1 PARTITION TYPES.................77 6.........1 MODULAR DEVELOPMENT ............................................................................................................................................................................................................................................................................4 COMPLEX FLAT FILE STAGE.....................................................8 2 STANDARDS.......................................................................................................81 7............................................................................................................................................................................................1 DIRECTORY STRUCTURES.....................................................................................

...............................133 12....................................124 11.3 MINIMIZING RUNTIME PROCESSES AND RESOURCE REQUIREMENTS.................................. JOIN VS....................................................................................134 APPENDIX A: STANDARD PRACTICES SUMMARY...............................................................................................................154 APPENDIX F: SORTING AND HASHING ADVANCED EXAMPLE...............................129 12....87 8.................................................................................................................................................................................................................................125 11...... stored in a retrieval system................................................................................................................................ MERGE................................. or translated into any language in any form by any means without the written permission of IBM................................95 10 DATABASE STAGE GUIDELINES............................................ transmitted.......................2 MODIFY STAGE.......94 9...................................................................................................91 9 COMBINING DATA.......................................................127 12 PERFORMANCE TUNING JOB DESIGNS..................................124 11..............................................................................................117 10................................................................................................................................................................................................................................................146 APPENDIX C: UNDERSTANDING THE PARALLEL JOB SCORE............4 VIEWING THE GENERATED OSH.............................................................................................................. transcribed...................................................................................................................................... 2006 5 of 179 © 2006 IBM Information Integration Solutions............................................................7 TERADATA DATABASE GUIDELINES....................148 APPENDIX D: ESTIMATING THE SIZE OF A PARALLEL DATA SET...................................2 CAPTURING UNMATCHED RECORDS FROM A JOIN........................................ ..................103 10.......................1 WARNING ON SINGLE-NODE CONFIGURATION FILES.............................................................................96 10...................................................................................................................5 ORACLE DATABASE GUIDELINES...................................................113 10............................................3 HOW TO ISOLATE AND DEBUG A PARALLEL JOB...............2 DEBUGGING ENVIRONMENT VARIABLES....................................................119 10....................................4 ODBC ENTERPRISE GUIDELINES....1 LOOKUP VS........ All rights reserved.....................2 UNDERSTANDING OPERATOR COMBINATION.................129 12........................................................................................................1 DATABASE DEVELOPMENT OVERVIEW................................................................................3 THE AGGREGATOR STAGE.......................................................................................................................................................................................................................................131 12..............................................126 11............................................4 UNDERSTANDING BUFFERING.......................................................................160 Parallel Framework Red Book: Data Flow Job Design July 17...........96 10...........................120 11 TROUBLESHOOTING AND MONITORING..............................................................................................................................................6 SYBASE ENTERPRISE GUIDELINES............................................Information Integration Solutions Center of Excellence 8.....114 10...........................1 HOW TO DESIGN A JOB FOR OPTIMAL PERFORMANCE........140 APPENDIX B: DATASTAGE NAMING REFERENCE...............................94 9.....................................5 INTERPRETING THE PARALLEL JOB SCORE..................3 INFORMIX DATABASE GUIDELINES................................................. No part of this publication may be reproduced.....1 TRANSFORMER STAGE..........................................................................................124 11............................................................................................................................94 9.......................................................................................................2 DB2 GUIDELINES....153 APPENDIX E: ENVIRONMENT VARIABLE REFERENCE.......................................

Here is the general flow diagram for Data Stage Enterprise Edition jobs: Before Job Subroutine Halt on Error ? Yes Exit Failure No Create Reject Files (Limited) Read Input Data Halt on Error ? Yes Exit Failure No Create Error and Reject Files Perform Validations Errors and Warnings Halt on Error ? Yes Exit Failure No Create Error and Reject Files Perform Transformations Halt on Error ? Yes Exit Failure No Create Reject Files (Limited) Perform Load and/or Create Intermediate Datasets Over Job W arning Threshold? Yes Exit Failure No Parallel Framework Red Book: Data Flow Job Design July 17. databases and DS/EE Data Sets. and • Provisioning data targets. . stored in a Subroutine retrieval system. All rights reserved. transcribed. Performing row validation to support data quality. Performing transformation from data sources to data targets. transmitted. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. 2006 6 of 179 After Job © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence 1 Data Integration Overview Work performed by Data Integration jobs fall into 4 general categories: • • • Reading input data including sequential files.

transmitted. In most production deployments. Cron. Parallel Framework Red Book: Data Flow Job Design July 17. auditing/capture. 2006 7 of 179 © 2006 IBM Information Integration Solutions. and together form a single end-to-end module within a DataStage application. etc). No part of this publication may be reproduced. error logging. stored in a retrieval system. built from individual Parallel jobs assembled in DataStage Enterprise Edition (“DS/EE”) controlled as modules from master DataStage Sequence jobs. Management. . and Production Automation. as illustrated in the example below: These job Sequences control the interaction and error handling between individual DataStage jobs.Information Integration Solutions Center of Excellence 1. This provides a level of granularity and control that is easy to manage and maintain. transcribed. Job sequences also provide the recommended level of integration with external schedulers (such as AutoSys.1 Job Sequences As shown in the previous diagram. These topics are discussed in Parallel Framework Standard Practices: Administration. and provides an appropriate leveraging of the respective technologies. etc). CA7. or translated into any language in any form by any means without the written permission of IBM. ETL development is intended to be modular. All rights reserved. Job Sequences require a level of integration with various production automation technologies (scheduling.

. Non-reference data or independent data are candidates. No part of this publication may be reproduced. and preserves the compute effort of long running transformation jobs. Hybrid Provisioning Data must NOT be changed by any method unless jobs transforming an entire subject area have successfully completed. Any target where either all sources have been successfully transformed or where the resources required to transform the data must be preserved in the event of a load failure or where the provisioning will take so long that it increases the probability of job failure if the job includes transformation and provisioning. stored in a retrieval system. and Hybrid jobs do both. Transformation jobs prepare data for provisioning jobs Provisioning jobs load transformed data. 2006 8 of 179 © 2006 IBM Information Integration Solutions. Data can be changed regardless of success or failure. or translated into any language in any form by any means without the written permission of IBM. transmitted. and Provisioning. transcribed. All rights reserved. This prevents partial replacement of reference data in the event of transformation failure. or where the resource requirements for data transformation are very large. The following table defines when each type should be used: Type Transformatio n Data Requirements Data must NOT be changed by any method unless jobs transforming an entire subject area have successfully completed. Parallel Framework Red Book: Data Flow Job Design July 17. Neither the transformation nor provisioning requirements are large. Example Reference tables upon which all subsequent jobs and/or the current data target (usually a database) will depend. The data target (usually a database) must allow subsequent processing of error or reject rows and tolerate partial or complete non-update of targets. or long running provisioning processes.2 - Job Types Nearly all data integration jobs fall into three major types: Transformation.Information Integration Solutions Center of Excellence 1. or where the resource requirements for data provisioning are very large. Hybrid.

transcribed. stored in a retrieval system. The following example transformation job demonstrates the use of write-through cache DS/EE Data Sets: The target table is among the inputs. some of which may be write-through cache Data Sets. data sources. . 2006 9 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. are processed to produce a load-ready Data Set that represents either the entire target table or new records to be appended to the target table. No part of this publication may be reproduced.1 Transformation Jobs In transformation jobs. that Data Set qualifies as write-through cache and may be used as source data instead of the target table. If the entire target table is held in the load-ready Data Set.2. All rights reserved. transmitted.Information Integration Solutions Center of Excellence 1. Parallel Framework Red Book: Data Flow Job Design July 17.

or translated into any language in any form by any means without the written permission of IBM. transmitted. All rights reserved. . No part of this publication may be reproduced. 2006 10 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. transcribed.Information Integration Solutions Center of Excellence The following example transformation job does NOT produce write-through cache – its sources do NOT include the target table. Parallel Framework Red Book: Data Flow Job Design July 17.

2 Hybrid Jobs The following example hybrid job demonstrates several interesting techniques that might be used in more complex jobs. This job also loads the target database table and creates write-through cache.Information Integration Solutions Center of Excellence 1. The non-validated records. and described below following the highlighted areas from Left to Right: A column generator inserts the key column for a join and generates a single value guaranteed to never appear in the other input(s) to the join. No part of this publication may be reproduced.2. Parallel Framework Red Book: Data Flow Job Design July 17. it is rejected by the transformer to a reject port and the validation is not performed for those records. The key column for a Referential Integrity check is validated by a Transformer stage. . the cache is deleted. transmitted. producing an ordered and linked associative table. and the write-through cache records from the last load of the target database are merged. we replicated the Oracle structure (lower input) for each country found in the write-through cache country dataset (upper input). forcing other jobs that might depend on this data to access the existing (not updated) target database table. Some of the more interesting solutions in this job are circled. or translated into any language in any form by any means without the written permission of IBM. The de-duplicated records are re-grouped and ordered before calculation of the terminating keys. This enforces a coherent view of the subject area from either cache (current state if all jobs complete successfully) or target tables (previous state if any job fails). stored in a retrieval system. 2006 11 of 179 © 2006 IBM Information Integration Solutions. If the key column is NULL. the validated records. transcribed. All rights reserved. if the load fails. In this case. The merged records are grouped and ordered before being de-duplicated to remove obsolete records. In this case. By specifying a full-outer join we produce a Cartesian product dataset.

3 Provisioning Jobs This example provisioning job demonstrates the straightforward approach to simple provisioning tasks. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 12 of 179 © 2006 IBM Information Integration Solutions. transmitted. .2. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. stored in a retrieval system. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 1.

All rights reserved.1 Directory Structures 2. .. /Ascential /Scratch0 /ScratchN /Project_A /patches /P /DataStage /Project_Z . 2. Figure 1: RecommendedGigabyte 1 DataStage Install.. File systems are highlighted in blue /Project_Z .. and Data Directories /Project_Z Parallel Framework Red Book: Data Flow Job Design July 17.. transcribed. and Project Directory Structures The following diagrams depict the IBM WebSphere DataStage software directory structures and the support directory structures.. or translated into any language in any form by any means without the written permission of IBM. Development standards can also make it easier to integrate external processes such as automated auditing and reporting..1 Data.1. These directories are configured during product installation. 2006 13 of 179 /Projects / /Project_A © 2006 IBM Information Integration Solutions.. and to build technical and support documentation. Install.. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 2 Standards Establishing consistent development standards helps to improve developer productivity and reduce ongoing maintenance costs. /DataN /P .. Scratch.. transmitted. stored in a retrieval system. Install FS Install File System Scratch File Systems . / /DSEngine Data File Systems /PXEngine /Data0 /Configurations /Project_A .

the DataStage Administrator client creates its projects (repositories) in the Projects directory of the DataStage installation directory. In general.. For this reason. All rights reserved. as disk space is typically limited in production install directories. consider creating separate file systems for each Scratch and Data resource partition. stored in a retrieval system. projects should be installed in their own file system. In order to scale I/O for DataStage. sort memory overflow. it is possible to create separate file systems at non-root levels. the administrator should consider creating separate file systems for each Scratch and Resource partition. . Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. transcribed.Information Integration Solutions Center of Excellence By default./Datasets The DataStage Administrator should ensure that these default directories are never used by any parallel configuration files. NOTE: On some operating systems. As a standard practice. No part of this publication may be reproduced. It is a bad practice to share the DataStage project file system and conductor file system with volatile files like scratch files and Parallel data set part files./Scratch $DSHOME/.. because they increase the risk of filling the DataStage project file systems. The DataStage installation creates the following two directories: $DSHOME/. as a separate file system for the Projects sub directory within the DataStage installation. transmitted. This best practice advocates creating subdirectories for each project for each scratch and disk partition. it is a bad practice to create DataStage projects in the default directory. Consider naming the file systems in accordance with partition numbers in your DataStage EE Configuration file. To scale I/O performance within DataStage. 2006 14 of 179 © 2006 IBM Information Integration Solutions. Scratch is used by the EE framework for temporary files such as buffer overflow. This is illustrated in the above diagram.

target data files. location of source data files. All rights reserved. /Project_Z Within the separate Staging file system. qa. If the file system is not /archive /archive /archive shared across multiple servers. error and reject files.. stored in a retrieval system. Within each deployment directory. No part of this publication may be reproduced.. . data directories are implemented for each deployment phase of a job (development../archive of these development phases may be present on a local file not all system. files are separated by Project name as shown below. or translated into any language in any form by any means without the written permission of IBM. /Project_Z July 17.. .. system integration. transmitted.Information Integration Solutions Center of Excellence /Staging /dev /si /qa /prod /Project_A /Project_A /Project_A /Project_A /archive /archive /archive /archive . 2006 15 of 179 . /Project_A /archive /si /qa /prod Parallel Framework Red Book: Data Flow Job Design Figure 2: DataStage Staging Directories /Project_Z /Project_Z © 2006 IBM Information Integration Solutions.. subdirectory created for each project location of compressed archives created by archive process of previously processed files System Integration (also known as “test”) data tree Quality Assurance data tree Production data tree . transcribed.. and production) as appropriate. /Staging /dev Top-Level Directory development data tree..

/Project_Z . or translated into any language in any form by any means without the written permission of IBM.. Project_Plus Project_Plus Directory Hierar /si /qa /prod /dev /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /sql /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /Project_A /bin /src /doc /datasets /logs /params /schemas /scripts /sql /Project_A /bin /s /d /datase /logs /param /schema /scripts /sql Figure 3: Project_Plus/sql Directory Structure Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. such as the operating system.1. which will enable DataStage applications to move through the life cycle without any code changes.Information Integration Solutions Center of Excellence 2. . 16 of 179 . or it could require scripts for example integrating with an Enterprise Scheduler. This will provide a convenient vehicle to group and manage resources used by a project.. To completely integrate all aspects of a DataStage application the directory structure that is used for integration with external entities should be defined in a way that provides a complete and separate structure in the same spirit as a DataStage project. All rights reserved. stored in a No part of this /Project_Z /Project_Z retrieval system....2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities. another application or middle ware. July 17. The integration can be as simple as a file system for housing source files. transmitted... 2006 . The directory structure will be made transparent to the DataStage application.. . transcribed. through the use of environment variables./Project_Z publication may be reproduced. A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project. Environment variables are a critical portability tool.

a copy of dsenv and copies of DSParams. No part of this publication may be reproduced. . If the file system is not shared across multiple servers. or translated into any language in any form by any means without the written permission of IBM. transcribed. system integration. and shells location of documentation for programs found in /bin subdirectory location of source code and makefiles for items found in /bin subdirectory (Note: depending on change management policies. and production) as appropriate. for example the following diagram shows Project_Plus variables as defined in the DataStage Administrator. DataStage routines. stored in a retrieval system. 2006 17 of 179 © 2006 IBM Information Integration Solutions. not all of these development phases may be present on a local file system.$ProjectName project files location of Orchestrate schema files location of operating system (shell) script files location of maintenance or template SQL system integration (aka “test”) code tree quality assurance code tree production code tree /datasets /logs /params /schemas /scripts /sql /si /qa /prod In support of a Project_Plus directory structure environment variable parameters should be configured. this directory may only be present in the /dev development code tree) location of DataSet header files (. directories are created for each deployment phase of a job (development.Information Integration Solutions Center of Excellence Within the Project_Plus hierarchy. Parallel Framework Red Book: Data Flow Job Design July 17. utilities. transmitted.ds file) location of custom job logs and reports location of parameter files for automated program control. Project_Plus /dev /Project_A /bin /doc /src Top-Level of Directory Hierarchy development code tree subdirectory created for each project location of custom programs. All rights reserved. qa. BuildOps.

2006 18 of 179 © 2006 IBM Information Integration Solutions. or an XML extract Parallel Framework Red Book: Data Flow Job Design July 17. A consistent naming standard is essential to • maximize the speed of development • minimize the effort and cost of downstream maintenance • enable consistency across multiple teams and projects • facilitate concurrent development • maximize the quality of the developed application • increase the readability of the objects in the visual display medium • increase the understanding of components when seen in external systems. transcribed. the Designer tool assigns default names based on the object type. stored in a retrieval system.Information Integration Solutions Center of Excellence Figure 4: Project_Plus Environment Variables In some implementations. No part of this publication may be reproduced. DataStage offers (within certain restrictions) flexibility to developers when naming various objects and components used to build a data flow. for example all jobs are invoked with the same Script. and the order the item is placed on the design canvas. there may be external entities that are shared with other DataStage projects. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. transmitted. . for example in WebSphere MetaStage. nor do they adequately document the business rules or subject areas.2 Naming Conventions As a graphical development environment. By default. 2. While the default names may create a functional data flow. they do not facilitate ease of maintenance over time. A similar directory structure to the Project_Plus structure could be configured and referred to as DataStage_Plus.

while maintaining quality. Reference (Lookup). • Since much of the development work is done using a click. 2. Subject Modifier. In the case of a data store the class word will refer to the type of data store. the class word is used to identify either a type of object or the function that a particular type of object will perform. more effort can be put into analysis and design. transmitted. Any set of standards needs to take on the culture of an organization.Information Integration Solutions Center of Excellence This section presents a set of standards and guidelines to apply to developing data integration applications using DataStage Enterprise Edition. • There can be a much tighter link between design and development. Table. 2006 19 of 179 © 2006 IBM Information Integration Solutions.1 Key Attributes of the Naming Convention This naming convention is based on a three-part convention: Subject. Parallel Framework Red Book: Data Flow Job Design July 17. and many of these benefits were used to establish this naming standard: • With rapid development. All rights reserved. As an example. For example. a transformer might be named: Data_Block_Split_Tfm As a guideline. objects can be sub-typed (for example. the term “Standard” refers to those principles that are required. View. enabling a greater understanding of the requirements and greater control over how they are delivered. No part of this publication may be reproduced. There are a number of benefits from using a graphical development tool like DataStage. for example: Data Set.2. and so forth. In some cases where appropriate. . three or four letter abbreviation. but not required. drag and drop paradigm there is less typing involved hence the opportunity to use longer more meaningful. stored in a retrieval system. In these cases the class word represents the subtype. Where there is no sub classification required then the class word will simply refer to the object. while the term “Guideline” refers to recommended. transcribed. to be tuned according to needs. in the case of a link object. Moving or Writing data (or within a Sequence Job. and Class Word In the context of DataStage. so it is envisaged that these standards will develop and will adapt over time to suit both the organization and the purpose. Where it is a two letter abbreviation both letters should be capitalized. principles. the class word refers to the functions of Reading. the moving of a message). more readable names. Where it is a three or four letter abbreviation then it should be word capitalized. or translated into any language in any form by any means without the written permission of IBM. a Left Outer Join). Sequential File. Throughout this section. the Class Word is represented as a two.

so a pattern of work needs to be identified and adopted to help development. documentation is often something that is left until later. all word abbreviations should be referenced by the long form to get used to saying the name in full even if reading the abbreviation. over using the Prefix approach. a derivation. whatever tool you use.3 Documentation and Metadata Capture One of the major problems with any development effort. Though DataStage imposes some limitations on the type of characters and length of various object names. For stages with multiple links. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. This type of approach takes extra effort at first. 2. is maintaining documentation. a link. One benefit of using the Subject. the first letter of each Word should be capitalized. In WebSphere MetaStage. will be to separate words by an Underscore which will allow clear identification of each work in a name. where possible. inadequately carried out. expanding the icon border can significantly improve readability.2. is to enable two levels of sorting or grouping. . Establishing standards also eases use of external tools and Parallel Framework Red Book: Data Flow Job Design July 17. The key issue is readability.Information Integration Solutions Center of Excellence A list of frequently-used Class Word abbreviations is provided in 12. Subject Modifier. Class Word approach.2 Designer Object Layout The effective use of naming conventions means that objects need to be spaced appropriately on the DataStage Designer canvas. Where possible. When development is more or less complete. This can help make them more productive and makes their work more easily read. Though best intentions are always apparent. For the purposes of documentation. allowing sort either by subject matter or by object type. 2. All rights reserved. a stage. Like a logical name. transcribed. 2006 20 of 179 © 2006 IBM Information Integration Solutions. consideration should be made to provide DataStage developers with higher resolution screens as this provides them with more screen display real-estate.2. the object type is defined in a separate field. Secondly the class word approach enables sub-classification by object type to provide additional information. This should be enhanced by also using Word Capitalization. This is the same or similar information that would be carried in a prefix approach. This will help re-enforce wider understanding of the subjects. and so forth. attention should be given to the layout to enhance readability before it is handed over to versioning.there is a field that denotes whether the object is a column. The “Snap to Grid” feature of Designer can help improve development speed. however. transmitted. the standard. a job design. Carrying this information as a separate attributes enables the first word of the name to be used as the subject matter. the abbreviated form is used.2 DataStage Naming Reference. DataStage provides the ability to document during development with the use of meaningful naming standards (as outlined in this section). when creating the object. for example. stored in a retrieval system.4.

The suffix of a Project name should be used to identify Development (“Dev”). transmitted. Examples of Project naming where the project is single application focused are: • “Accounting Engine NAB Development” would be named: Acct_Eng_NAB_Dev • “Accounting Engine NAB Production” would be named: Acct_Eng_NAB_Prod Examples of Project naming where the project is multi-application focused are: • Accounting Engine Development or Acct_Engine_Dev • Accounting Engine Production or Acct_Engine_Prod 2. No part of this publication may be reproduced. are Alpha Numeric and can also contain both Spaces and Underscores. 2. allowing related objects to be grouped together.2 Category Hierarchy DataStage organizes objects in its repository by Categories. All rights reserved. dialog box fields that specify a new category have only one input area for defining the Category name. stored in a retrieval system. 2. DataStage enforces the top level Directory Structure for different types of Objects (for example. It may or may not have a one to one relationship with an organizations’ project of work. Routines.Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage. developers have the flexibility to define their own Directory or Category hierarchy.4. Therefore Directory names should be Word Capitalized and separated by either an underscore or a space. This factor often can cause terminology issues especially in teamwork where both business and developers are involved.2.4. 2006 21 of 179 © 2006 IBM Information Integration Solutions.2. Category Names can be long. Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”). Test (“Test”). For example. Below this level. it can contain alpha-numeric characters and it can contain underscores. as well as documentation and auditing. Shared Containers. Jobs.2. Table definitions…). However with the limit of 18 characters the name is most often composed of abbreviations.1 Projects Each DataStage Project is a standalone repository. .3 Job Category Naming Within Designer. and Production (“Prod”).2.4 Naming Conventions by Object Type 2. or translated into any language in any form by any means without the written permission of IBM. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17. which can provide impact analysis.4. transcribed. The name of a DataStage Project may only be 18 characters in length.

Jobs and Job Sequences are grouped together in the same scope as the technical design documents. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 5: Creating Category Hierarchies The main reason for having Categories is to group related objects. a Category level should only contain objects that are directly related. Categorization by Functional Module For a given application. 2006 22 of 179 © 2006 IBM Information Integration Solutions. jobs that read write-through cache for a ECRP subset in the ECRDEV project that cleanse and load multi-family mortgage data and are driven by a sequencer might have a hierarchy that looks like the following example: Parallel Framework Red Book: Data Flow Job Design July 17. For example. stored in a retrieval system. All rights reserved. not within a category. For example. or translated into any language in any form by any means without the written permission of IBM. . with sub-levels for individual functional modules. Within each functional module category. Note that Job names must be unique within a DataStage project. Organizing related DataStage objects within categories also facilitates backup/export/import/change control strategies for projects since Manager can import/export objects by category grouping. a job category might contain a Job Sequence and all the jobs and only those jobs that are contained in that sequence. Where possible. transcribed. transmitted. all Jobs and Job Sequences will be grouped in a single parent Category.

Once created. but this can be overridden during import. Table Definitions are always categorized using two level names. Remembering that Job names must be unique within a given project. Saved.).2. PlugIn. ECRP and Templates. and there are 2 additional high-level categories. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 6: Categorization by Functional Module Categorization by Developer In development projects. or translated into any language in any form by any means without the written permission of IBM. On import. 2. if these TableDefs are to be used by other jobs. the first level Table Definition category is identified as the “Data Source Type” and the second level categorization is referred to as the “Data Source Name” as shown in the example on the below. project manager. New Table Definition categories can be created within the repository by right-clicking within the Table Definitions area of the DataStage project repository and choosing the “New Category” command. etc. categories and metadata. transcribed. transmitted. 2006 23 of 179 © 2006 IBM Information Integration Solutions.. When implementing a customized Table Definition categorization. establishing a Table Definition categorization that matches project development organization is recommended. The placement of these fields varies with the method of metadata import. TableDefs that remain in the Saved category Parallel Framework Red Book: Data Flow Job Design July 17.4 Table Definition Categories Unlike other types of DataStage objects. All rights reserved. categories will be created for each developer as their personal sandbox and place they perform unit test activities on jobs they are developing. and the responsibility of the development manager assigned the DataStage Manager role to ensure that projects are not obese with unused jobs. DataStage assigns the level names based on the source of the metadata import (for example. It is the responsibility of each developer to delete unused or obsolete code. two developers cannot save a copy of the same job with the same name within their individual “sandbox” categories – a unique Job name must be given. Although the default table definition categories are useful from a functional perspective. they must be moved to the appropriate category and re-imported from that category in every job where they are used. Temporary TableDefs created by developers to assist with job creation appear under the Saved category by default.. care must be taken to override the default choices for category names during Table Definition import. By default. two developers have private categories for sandbox and development activities.4. In the previous illustration. Orchestrate. . stored in a retrieval system.

or translated into any language in any form by any means without the written permission of IBM.4. Parallel Framework Red Book: Data Flow Job Design July 17.2. No part of this publication may be reproduced.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters. An alternative implementation is to set the “Data source name” to that of the source system or schema. stored in a retrieval system. . e. All rights reserved. Job and Job Sequence names should be descriptive and should use word capitalization to make them readable. Examples of Job naming are: • CodeBlockAggregationJob • CodeBlockProcessingSeq Jobs should be organized under Category Directories to provide grouping such that a Directory should contain a Sequence Job and all the Jobs that are contained within that sequence. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category. transmitted. This will be discussed further in Section 2. Each subject area will have a master category. In this example.g. 2006 24 of 179 © 2006 IBM Information Integration Solutions. with sub-categories intended to identify the type of the source. transcribed. Figure 7: Table Definition Categories 2. numbers.Information Integration Solutions Center of Excellence should be deleted as soon as possible.: Datasets.4. Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”.2.2 Category Hierarchy. and underscores only.g: DWPH1 or ECRP. the TableDefs have been grouped into a master category of Custom. Because the name of can be long. e. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”.

7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores. 2. links represent the flow of a message from one activity / step to the next. so word capitalization should be used for readability. transcribed. stored in a retrieval system. .4. during Link Ordering). instead of using the default “DSLink#” (where “#” is an assigned number). transmitted.8 Links Within a DataStage Job. The class word suffix is “Parm”.2. To differentiate between Parallel Shared Containers and Server Shared Containers. a character code is automatically added to that instance of its use throughout the project. Furthermore. 2006 25 of 179 © 2006 IBM Information Integration Solutions. Within the graphical Designer environment.Information Integration Solutions Center of Excellence 2. or translated into any language in any form by any means without the written permission of IBM. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. links are objects that represent the flow of data from one stage to the next. When a Shared Container is used. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. Examples of Shared Container naming are: • AuditTrailPsc (this is the original as seen in the Category Directory) • AuditTrailPscC1 (This is an instance of use of the above shared container) • AuditTrailPscC2 (This is another instance of use of the same shared container) In the above examples the characters “C1” and the “C2” are automatically applied to the Shared Container Stage by DataStage Designer when dragged onto the design canvas.2.4. Therefore the parameter name must be made readable using Capitalized words separated by underscores. stage editors identify links by name.6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2.2. All rights reserved. It is particularly important to establish a consistent naming convention for link names. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17.4. having a descriptive link name reduces the chance for errors (for example. Within a Job Sequence. No part of this publication may be reproduced. the following Class Word naming is recommended: • Psc = Parallel (Enterprise Edition) Shared Container • Ssc = Server Edition Shared Container IMPORTANT: Use of Server Edition Shared Containers is discouraged within a parallel job. It is optional as to whether you decide to change this code to something meaningful.

Transformer. always specify a unique link name within a particular Job or Job Sequence by including a number. Database. No part of this publication may be reproduced. transcribed. . etc) • The type of movement may optionally be part of the Class Word. Parallel Framework Red Book: Data Flow Job Design July 17. Merge. establishing standardized link names makes it easier to understand results and audit counts. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. (The DataStage Designer does not require link names on different stages to be unique.) • Examples Link names: • Input Transactions: “lnk_Txn_In” • Reference Account Numbers: “lnk_Account_Ref” • Customer File Rejects: “lnk_Customer_Rej” • Reception Succeeded Message or “lnk_Reception_Succeeded_Msg” 2. 2006 26 of 179 © 2006 IBM Information Integration Solutions. the same name may be appropriate for multiple links.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. or 4 character abbreviation should be used for the Class Word suffix. a 2. for example: o “In” for input o “Out” for output o “Upd” for updates o “Ins” for inserts o “Del” for deletes o “Get” for shared container inputs o “Put” for shared container output As data is enriched through stages. In this case. the link name should include the link type (reference. Within a Job or Job Sequence. reject) to reinforce the visual cues of the Designer canvas: o “Ref” for reference links (Lookup) o “Rej” for reject links (Lookup. after the subject name and subject modifier. Instead of using the full object name. through Job reporting). transmitted. based on the order the object was added to the flow. The following rules can be used to establish a link name: • Use the prefix “lnk_” before the subject name to differentiate with stage objects • The link name should define the subject of the data that is being moved • For non-stream links. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12. stage names must be unique.4. These names are based on the type of stage (object) and a unique number.4. 3.2. stored in a retrieval system.Information Integration Solutions Center of Excellence example. Sequential File.2 DataStage Naming Reference.

Information Integration Solutions Center of Excellence

2.2.4.10 Data Stores For the purposes of this section, a data store is a physical piece of disk storage where data is held for some period of time. In DataStage terms, this can be either a table in a database structure or a file contained within a disk directory or catalog structure. Data held in a database structure is referred to as either a Table or a View. In data warehousing, two additional subclasses of table might be used: Dimension and Fact. Data held in a file in a directory structure will be classified according to its type, for example: Sequential File, Parallel Data Set, Lookup File Set, etc. The concept of source and target can be applied in a couple of ways. Every job in a series of jobs could consider the data it gets in to be a source and the data it writes out as being a target. However for the sake of this naming convention a Source will only be data that is extracted from an original system and Target will be the data structures that are produced or loaded as the final result of a particular series of jobs. This is based on the purpose of the project – to move some data from a source to a target. Data Stores used as temporary structures to land data between jobs, supporting restart and modularity, should use the same names in the originating job and any downstream jobs reading the structure. Examples of Data Store naming are: • Transaction Header Sequential File or Txn_Header_SF • Customer Dimension or Cust_Dim (This optionally could be further qualified as Cust_Dim_Tgt if you wish to qualify it as a final target) • Customer Table or Cust_Tab • General Ledger Account Number View or GL_Acctno_View 2.2.4.11 Transformer Stage and Stage Variables A Transformer Stage name can be long – over 50 characters and can contain underscores. Therefore the name can be descriptive and readable through word capitalization and underscores. DataStage Enterprise Edition supports two types of Transformers: • “Tfm”: Parallel (Enterprise Edition) Transformer • “BTfm”: BASIC (Server Edition) Transformer IMPORTANT: For maximum performance and scalability, BASIC Transformers should be avoided in Enterprise Edition data flows. A Transformer Stage Variable can have a long name consisting of alphanumeric characters but not underscores. Therefore the Stage Variable name must be made readable only by using Capitalized words. The Class Word suffix is Stage Variable or “SV”. Stage Variables should be named according to their purpose.

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

27 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

When developing Transformer derivation expressions, it is important to remember Stage variable names are case sensitive. 2.2.4.12 DataStage Routines DataStage BASIC routine names will indicate their function and they will be grouped in sub-categories by function under a main category of Custom, for example.: Routines/Custom/SetDSParamsFromFile. A How-To document describing the appropriate use of the routine must be provided by the author of the routine, and placed in a documentation repository. DataStage Custom Transformer routine names will indicate their function and they will be grouped in sub-categories by function under a main category of Custom, for example: Routines/Custom/DetectTeradataUnicode. Source code, a makefile, and the resulting object for each Custom Transformer routine must be placed in the project phase source directory, e.g.: /home/dsadm/dev/bin/source. 2.2.4.13 File Names Source file names should include the name of the source database or system and the source table name or copybook name. The goal is to connect the name of the file with the name of the storage object on the source system. Source flat files will have a unique serial number composed of the date, “_ETL_” and time, for example: Client_Relationship_File1_In_20060104_ETL_184325.psv. Intermediate datasets are created between modules. Their names will include the name of the module that created the dataset OR the contents of the dataset in that more than one module may use the dataset after it is written, for example: BUSN_RCR_CUST.ds Target output files will include the name of the target database or system, the target table name or copybook name. The goal is the same as with source files – to connect the name of the file with the name of the file on the target system. Target flat files will have a unique serial number composed of the date, “_ETL_” and time, for example: Client_Relationship_File1_Out_20060104_ETL_184325.psv Files and datasets will have suffixes that allow easy identification of the content and type. DataStage proprietary format files have required suffixes and are identified in italics in the table below which defines the types of files and their suffixes. File Type
Parallel Framework Red Book: Data Flow Job Design

Suffix
July 17, 2006 28 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Flat delimited and non-delimited files Flat pipe (|) delimited files Flat comma-and-quote delimited files DataStage datasets DataStage filesets DataStage hash files Orchestrate schema files Flat delimited or non-delimited REJECT files DataStage REJECT datasets Flat delimited or non-delimited ERROR files DataStage ERROR datasets Flat delimited or non-delimited LOG files .dat. .psv .csv. .ds. .fs .hash. .schema. .rej. _rej.ds. .err. _err.ds. .log.

2.3

Documentation and Annotation

DataStage Designer provides description fields for each object type. These fields allow the developer to provide additional descriptions that can be captured and used by administrators and other developers. The Short Description field is also displayed on summary lines within the Director and Manager clients. At a minimum, description annotations must be provided in the Job Properties Short Description field for each job and job sequence, as shown below:

Figure 8: Job Level Short Description Within a job, the Annotation tool should be used to highlight steps in a given job flow. Note that by changing the vertical alignment properties (for example, Bottom) the annotation can be drawn around the referenced stage(s), as shown in the following example.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 29 of 179
© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Examples of such annotations include: Job “short” description: This Job takes the data from GBL Oracle Table AD_TYP and does a truncate load into Teradata Table AD_TYP. No part of this publication may be reproduced. transcribed. transmitted.RcR_GLOBAL_BUSN_CAT_TYP table from jpORACLE_SERVER using the ODBC driver.Information Integration Solutions Center of Excellence Figure 9: Example Job Annotation DataStage also allows descriptions to be attached to each stage within the General tab of the stage properties. Each stage should have a short description of its function specified within the stage properties. All rights reserved. More complex operators or operations should have correspondingly longer and more complex explanations on this tab. These descriptions will appear in the job documentation automatically generated from jobs and sequencers adhering to the standards in this document. ODBC Enterprise stage read: Read the GLO. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. . There are no selection criteria in the WHERE clause. 2006 30 of 179 © 2006 IBM Information Integration Solutions. There are no selection criteria in the WHERE clause. Parallel Framework Red Book: Data Flow Job Design July 17. Oracle Enterprise stage read: Read the GLOBAL.GLOBAL_REST_CHAR table from jpORACLE_SERVER using the Oracle Enterprise operator.

This stage validates the input and continues. No part of this publication may be reproduced. This stage renames and/or drops columns and is NOT optimized out. etc. . Teradata MultiLoad stage: Load the RcR_GLOBAL_LCAT_TYP table. 2. or translated into any language in any form by any means without the written permission of IBM. This stage identifies changes and drops records not matched (not updated).4 Working with Source Code Control Systems DataStage’s built-in repository manages objects (jobs. 2006 31 of 179 © 2006 IBM Information Integration Solutions. This is the target file for business qualification process rejects. Parallel Framework Red Book: Data Flow Job Design July 17. this repository is not capable of managing nonDataStage components (for example. job scheduler configurations. However. This stage is cosmetic and is optimized out. custom components) during job development. UNIX shell scripts. sequences. Data Set stage: This stage writes the GLOBAL_Ad_Typ dataset which is used as write-through cache to avoid the use of Teradata in subsequent jobs. routines. transcribed. All rights reserved. environment files. and to a dataset for use as write-through cache. Transformer stage: This stage generates sequence numbers that have a less-than file scope.) that may be part of a completed application. Copy stage This stage sends data to the TDMLoadPX stage for loading into Teradata. Sequential file stage: This is the source file for the LANG table. This stage converts null dates. which is used as write-through cache to avoid the use of Teradata. This stage reads the GLOBAL_Lcat dataset. stored in a retrieval system.Information Integration Solutions Center of Excellence Remove Duplicates stage This stage removes all but one record with duplicate BUSN_OWN_TYP_ID keys. Modify stage: This stage performs data conversions not requiring a transformer. table definitions. Lookup stage This stage validates the input and writes rejects. transmitted.

All rights reserved. . Designer facilitates iterative job design.4. SCCS) are useful for managing the development lifecycle of all components of an application. Typically. DataStage does not directly integrate with source code control systems. operating system backups of the project repository files only establish a “point in time”. The Manager client is the primary interface to the DataStage object repository. rules should be defined for when this transfer should take place. Parallel Framework Red Book: Data Flow Job Design July 17. organized into specific releases for version control. 2006 32 of 179 © 2006 IBM Information Integration Solutions. Grouping related objects by folder can simplify the process of exchanging information with the external source code control system.1 Source Code Control Standards The first step to effective integration with source code control systems is to establish standards and rules for managing this process: a) Establish Category naming and organization standard DataStage objects can be exported individually or by category (folder hierarchy). when a set of objects has completed initial development. Unless there is a need to parse information in the export file. it is important that an identified individual maintains backup copies of the important job designs using . milestone points on the development lifecycle are a good point for transferring objects to the source code control system . For these reasons. etc.Information Integration Solutions Center of Excellence Source code control systems (such as ClearCase. and so on. unit test. b) Define rules for exchange with source code control As a graphical development environment. stored in a retrieval system. and cannot be used to restore individual objects. Rather. you can export objects (job designs. although the XML file is generally much larger.DSX (DataStage eXport format) or . Both formats contain the same information.) from the repository as clear-text format files. The export file format for DataStage objects can be either . Using Manager.for example. No part of this publication may be reproduced.XML. This object grouping also helps establish a manageable “middle ground” between an entire project exports and individual object exports. It would be cumbersome to require the developer to check-in every change to a DataStage object in the external source code control system. 2. but it does offer the ability to exchange information with these systems. These files can then be checked into the external source code control system. Furthermore.DSX is the recommended export format. or translated into any language in any form by any means without the written permission of IBM.DSX file exports to a local or (preferably) shared file system. transmitted. they would not be an effective backup strategy. c) Don’t rely on the source code control system for backups Because the rules defined for transfer to the source code control system will typically be only at milestones in the development cycle. PVCS. custom stage types. . table definitions. user-defined routines. It is the responsibility of the DataStage developer to maintain DataStage objects within the source code system. transcribed.

These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide. The DataStage Manager can export at the Project. or by the individual DataStage developer. 2.2 Category Hierarchy. establishing and following a consistent naming and categorization standard is essential to the change management process. or translated into any language in any form by any means without the written permission of IBM. In either case. and individual Object levels. All rights reserved. All exports from the DataStage repository are performed on the Windows workstation. The DataStage client includes Windows command-line utilities for automating the export process. stored in a retrieval system. No part of this publication may be reproduced. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu. July 17.4.Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process. transcribed. • Select the object or category in the Manager browser.4. 2. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Category. .4. There is no server-side project export facility. Assigning related objects to the same category provides a balanced level of granularity when exporting and importing objects with external source code control systems. as explained in this section. the developer should create a local backup prior to implementing any extensive changes.2.2 Using Object Categorization Standards As discussed in Section 2. transmitted. It can be done interactively by the developer or project manager using the Manager client.

All rights reserved. Figure 11: Manager Export Options To export a group of objects to a single export file. 2006 34 of 179 © 2006 IBM Information Integration Solutions. Import can be interactive through the Manager client (as described in this section). the option “Selection: By category” should be specified in the “Options” tab. Parallel Framework Red Book: Data Flow Job Design July 17. The filename for export is specified in the “Export to file:” field at the top of the Export dialog. or translated into any language in any form by any means without the written permission of IBM. or automated through command-line utilities.4. Make sure all objects are saved and closed before exporting. check-in the exported .Information Integration Solutions Center of Excellence NOTE: Objects cannot be exported from DataStage if they are open in Designer. • Using your source code control utilities. . make sure the “Job Executables” category is checked.4 Import from Source Code Control System In a similar manner. the import of objects from an external source code control system is a straightforward process. transcribed. No part of this publication may be reproduced. stored in a retrieval system.DSX file 2. transmitted. If you wish to include compiled Transformer objects for a selected job.

Select the file you checked out of your source code control system by clicking on the ellipsis (“…”) next to the filename field in the import dialog.5 Understanding a Job’s Environment DataStage Enterprise Edition provides a number of environment variables to enable / disable product features and to fine-tune job performance. . After selecting your file. then compile the imported objects from Designer. • • Use the source code control system to check-out (or export) the . If the objects were not exported with the “Job Executables”. command-line import utilities are available for both Windows workstation and DataStage server platforms. click OK to import.DSX file to your client workstation. Although operating system environment variables can be set in multiple places. 2006 35 of 179 © 2006 IBM Information Integration Solutions. Choose “Import DataStage Components” from the “Import” menu. dsrpcd is started when the server installed. or using the Multi-Job Compile tool. Figure 12: Manager Import options • • The import of the . and should start Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. there is a defined order of precedence that is evaluated when a job’s actual environment is established at runtime: 1) The daemon for managing client connections to the DataStage server engine is called dsrpcd. The Windows workstation utilities (dsimport and dscmdimport) are documented in the DataStage Manager Guide.DSX file will place the object in the same DataStage category it originated from. All rights reserved. transcribed. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Unlike the export process. transmitted. stored in a retrieval system. it is possible to import the job executables from the DataStage server host using the dsjob command-line. as documented in the DataStage Development Kit chapter of the Parallel Job Advanced Developer’s Guide.DSX file using Manager. This means that if necessary it will create the Job Category if it doesn't already exits. Import objects in the . 2. By default (in a root installation). For test and production environments. Note that using dsjob will only import job executables job designs can only be imported using the Manager client or the dsimport or dscmdimport client tools.

on UNIX platforms is set in the etc/profile and $DSHOME/dsenv scripts. during job development. 3) Within Designer. transcribed. All rights reserved. the Designer parses the corresponding parallel configuration file to obtain a list of node maps and constraints (advanced stage properties). Typically this is used to pickup values set in the operating system outside of DataStage. Note that client connections DO NOT pick up per-user environment settings from their $HOME/. and Production Automation Best Practice for additional details. Any job-level settings for a specific environment variable will override any settings inherited from dsrpcd or from project-level defaults. Any project-level settings for a specific environment variable will override any settings inherited from dsrpcd. there are three special values that can be used for environment variables within job parameters: • causes the value of the named environment variable to be retrieved from the operating system of the job environment. On USS environments. Management. 2006 36 of 179 © 2006 IBM Information Integration Solutions. it is important to note that project-level environment variable settings are not exported when a project is exported. . These settings are stored in a file named DSPARAMS in the project directory. No part of this publication may be reproduced. dsrpcd can also be manually started and stopped using the $DSHOME/uv – admin command.) By default. (For more information. DataStage jobs inherit the dsrpcd environment. $PROJDEF Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. transmitted.profile script. the default DataStage environment is defined in the registry. Any project-level environment variables must be set for new projects using the Administrator client. environment variables may be defined for a particular job using the Job Properties dialog box. or by carefully editing the DSPARAMS file within the project. • causes the project default value for the environment variable (as shown on the Administrator client) to be picked up and used to set the environment variable and job parameter for the job. 2) Environment variable settings for particular projects can be set in the DataStage Administrator client. which. Refer to the DataStage Administration. To avoid hard-coding default values for job parameters. $ENV NOTE: $ENV should not be used for specifying the default $APT_CONFIG_FILE value because.Information Integration Solutions Center of Excellence whenever the machine is restarted. IMPORTANT: When migrating projects between machines or environments. or translated into any language in any form by any means without the written permission of IBM. the dsrpc environment is not inherited since DataStage jobs do not execute on the conductor node. On Windows. see the DataStage Administrator Guide.

It may be helpful to create a Job Template and include these environment variables in the parameter settings. No part of this publication may be reproduced.5.5. stored in a retrieval system. This variable should be included in all job parameters so that it can be easily changed at runtime. . but part of every job design so that it can be easily enabled for debugging purposes. $UNSET 2. and to document a few that are not part of the documentation.2 Understanding the Parallel Job Score) Includes a copy of the generated osh in the job’s DataStage log Outputs record counts to the DataStage job log as each operator completes processing. This setting should be disabled by default. transmitted.1. Does not report PIDs of DataStage “phantom” processes started by Server shared containers. Environment Variable $APT_CONFIG_FILE Setting filepath 1 $APT_DUMP_SCORE $OSH_ECHO $APT_RECORD_COUNTS 1 0 Description Specifies the full pathname to the EE configuration file. All rights reserved. This section is intended to call attention to some specific environment variables.4.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. Outputs actual runtime metadata (schema) to DataStage job log. Several environment variables are evaluated only for their presence in the environment (for example. and Data Sets. transcribed. APT_SORT_INSERTION_CHECK_ONLY). Maximum buffer delay in seconds $OSH_PRINT_SCHEMA 0 $APT_PM_SHOW_PIDS 1 $APT_BUFFER_MAXIMUM_TIMEOUT 1 Parallel Framework Red Book: Data Flow Job Design July 17. This setting should be disabled by default. providing detailed information about actual job flow including operators. processes. The count is per operator per partition.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment. or translated into any language in any form by any means without the written permission of IBM.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. 2006 37 of 179 © 2006 IBM Information Integration Solutions. or may be set on an individual basis within the properties for each job. Extremely useful for understanding how a job actually ran in the environment. Outputs EE score dump to the DataStage job log. These settings can be made at the project level. but part of every job design so that it can be easily enabled for debugging purposes. (see 12. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. 2.

2006 38 of 179 © 2006 IBM Information Integration Solutions. assisting in debugging.1. a number of environment variables will be mentioned for tuning the performance of a particular job flow. No part of this publication may be reproduced. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. The environment variables mentioned in this document are summarized in Appendix D: Environment Variable Reference.1A platforms only: On Tru64 platforms.Information Integration Solutions Center of Excellence On Solaris platforms only: When working with very large parallel Data Sets (where the individual data segment files are larger than 2GB). the environment variable $APT_PM_NO_SHARED_MEMORY should be set to 1 to work around a performance issue with shared memory MMAP operations. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. transmitted. . 2.5. All rights reserved.2 Additional Environment Variable Settings Throughout this document. you must define the environment variable $APT_IO_NOMAP On Tru64 5. or changing the default behavior of specific Enterprise Edition stages. This setting instructs EE to use named pipes rather than shared memory for local data transport. transcribed.

password. complex job that satisfies given functional requirements. Parallel Framework Red Book: Data Flow Job Design July 17. While it may be possible to construct a large. Factors to consider when establishing job boundaries include:  Establishing job boundaries through intermediate Data Sets creates “checkpoints” that can be used in the event of a failure when processing must be restarted. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously.2 Establishing Job Boundaries It is important to establish appropriate job boundaries when developing with DS/EE. It is for these reasons that long-running tasks are often segmented into separate jobs in an overall sequence. these standard parameters and settings should be made part of a Designer Job Template. create re-usable components such as parallel shared containers to encapsulate frequently-used logic. Where possible. To ease re-use. it is generally a good idea to land data to a parallel Data Set before loading to a target database unless the data volume is small or the overall time to process the data is minimal. . o For example. etc) and directories where files are stored. No part of this publication may be reproduced. • • • 3. For example. stored in a retrieval system. In some cases.Information Integration Solutions Center of Excellence 3 Development Guidelines 3. But functional requirements may not be the only factor driving the size of a given DataStage job. transmitted. functional requirements may dictate job boundaries. this may not be appropriate. and so forth. it may be appropriate to update all dimension values before inserting new entries in a data warehousing fact table. Create a standard directory structure outside of the DataStage project directory for source and target files. if the extract of source data takes a long time (such as an FTP transfer over a wide area network) it would be good to land the extracted source data to a parallel data set before processing. intermediate work files. Without these checkpoints. or translated into any language in any form by any means without the written permission of IBM. o As another example. A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN. user.1 Modular Development Modular development techniques should be used to maximize re-use of DataStage jobs and components: • Job parameterization allows a single job design to process similar logic instead of creating multiple copies of the same job. processing must be restarted from the beginning of the job flow. All rights reserved. transcribed. 2006 39 of 179 © 2006 IBM Information Integration Solutions.

Information Integration Solutions Center of Excellence

Larger, more complex jobs require more system resources (CPU, memory, swap) than a series of smaller jobs, sequenced together through intermediate Data Sets. Resource requirements are further increased when running with a greater degree of parallelism specified by a given configuration file. However, the sequence of smaller jobs generally requires more disk space to hold intermediate data, and the speed of the I/O subsystem can impact overall end-to-end throughput. Section 12.3: Minimizing Runtime Processes and Resource Requirements provides some recommendations for minimizing resource requirements of a given job design, especially when the volume of data does not dictate parallel processing.

Breaking large job flows into smaller jobs may further facilitate modular development and reuse if business requirements for more than one process depend on intermediate data created by an earlier job. The size of a job directly impacts the speed of development tasks such as opening, saving, and compiling. These factors may be amplified when developing across a wide-area or high-latency network connection. In extreme circumstances this can significantly impact developer productivity and ongoing maintenance costs. The startup time of a given job is directly related to the number of stages and links in the job flow. Larger more complex jobs require more time to startup before actual data processing can begin. Job startup time is further impacted by the degree of parallelism specified by the parallel configuration file. Remember that the number of stages in a parallel job includes the number of stages within each shared container used in a particular job flow.

As a rule of thumb, keeping job designs to less than 50 stages may be a good starting point. But this is not a hard-and-fast rule. The proper job boundaries are ultimately dictated by functional / restart / performance requirements, expected throughput and data volumes, degree of parallelism, number of simultaneous jobs and their corresponding complexity, and the capacity and capabilities of the target hardware environment. Combining or splitting jobs is relatively easy, so don't be afraid to experiment and see what works best for your jobs in your environment.

3.3

Job Design Templates

DataStage Designer provides the developer with re-usable Job Templates, which can be created from an existing Parallel Job or Job Sequence using the “New Template from Job” command.

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

40 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Template jobs should be created with: - standard parameters (for example, source and target file paths, database login properties…) - environment variables and their default settings (as outlined in Section 2.5.1 Environment Variable Settings) - annotation blocks In addition, template jobs may contain any number of stages and pre-built logic, allowing multiple templates to be created for different types of “standardized” processing. By default, the Designer client stores all job templates in the local “Templates” directory within the DataStage client install directory, for example, C:\Program Files\Ascential\DataStage751\Templates To facilitate greater re-use of job templates, especially in a team-based development, the template directory can be changed using the Windows Registry Editor. This change must be made on each client workstation, by altering the following registry key:
HKEY_LOCAL_MACHINE\SOFTWARE\Ascential Software\DataStage Client\CurrentVersion\Intelligent Assistant\Templates

3.4

Default Job Design

Default job designs include all of the capabilities detailed Section 2: Standards. Template jobs should contain all the default characteristics and parameters the project requires. These defaults provide at a minimum: 1. 2. 3. 4. 5. Development phase neutral storage (e.g.: dev, si, qa and prod); Support for Teradata, Oracle, DB2/UDB and SQL Server login requirements; Enforced project standards; Optional operational metadata (runtime statistics) suitable for loading into a database; and Optional auditing capabilities.

The default job design specifically will support the creation of write-through cache in which data in load-ready format is stored in DS/EE Data Sets for use in the load process or in the event the target table becomes unavailable. The default job design incorporates several features and components of DataStage that are used together to support tactical and strategic job deployment. These features include: 1. Re-start-able job sequencers which manage one or more jobs, detect and report failure conditions, provide monitoring and alert capabilities and support checkpoint restart functionality. 2. Custom routines written in DataStage BASIC (DS Basic) that detect external events, manage and manipulate external resources, provide enhanced notification and alert capabilities and interface to the UNIX operating system. 3. DataStage Enterprise Edition (DS/EE) ETL jobs that exploit job parameterization, runtime UNIX environment variables, and conditional execution.
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 41 of 179
© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

Each subject area is broken into sub-areas and each sub-area may be further subdivided. These sub-areas are populated by a DataStage job sequencer utilizing 2 types of DataStage jobs at a minimum: 1. A job that reads source data and • Transforms it to load-ready format • Optionally stores its results in a write-through cache DataStage Data Set or loads the data to the target table. 2. A job that reads the DataStage dataset and loads it to the target table. Other sections will discuss in detail each of the components and give examples of their use in a working example job sequencer.

3.5

Job Parameters

Parameters are passed to a job as either DataStage job parameters or as environment variables. Job parameters can be set from a file and are distinguished by the presence of a ‘jp’ prefix to the variable name. This prefix is part of the DataStage development standard. The names of environment variables have no prefix when they are set (UNIX_VAR=”some value”) and a prefix of “$” when used (myval=$UNIX_VAR). Job parameters are passed from a job sequencer to the jobs in its control as if a user were answering the runtime dialog questions displayed in the DataStage Director job-run dialogue. Default environment variables cannot be reset during this dialog unless explicitly specified in the job. The scope of a parameter depends on their type. Essentially: o The scope of a job parameter is specific to the job in which it is defined and used. Job parameters are stored internally within DataStage for the duration of the job, and are not accessible outside that job. o The scope of a job parameter can be extended by the use of job sequencer, which can manage and pass job parameters among jobs. o The scope of an environment variable is wider, as it is defined at operating system level, though conversely the use of environment variables is limited within this exercise. Job parameters are required for the following DataStage programming elements: 1. File name entries in stages that use files or Data Sets must NEVER use a hard-coded operating system pathname. a. Staging area files must ALWAYS have pathnames as follows: /jpSTAGING/jpENVIRON/jpSUBJECT_AREA[filename.suffix] b. DataStage datasets ALWAYS have pathnames as follows: /jpDSTAGE_ROOT/jpENVIRON/datasets/[filename.suffix]

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

42 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Database stages must ALWAYS use variables for the server name. The presence of rejects may indicate that a job has failed and prevent further processing. and further. transmitted. any additional columns are passed through the container at runtime without the need to separate and remerge. Reject files will include those records rejected from the ETL stream due to Referential Integrity failures. Use and management of job parameters. Error files will include those records from sources that fail quality tests.7 Error and Reject Record Handling Reject rows are those rows that fail active or passive business rule driven validation as specified in the job design document. schema (if appropriate). Using RCP. Parallel Framework Red Book: Data Flow Job Design July 17. The exact policy for each reject is specified in the job design document.Information Integration Solutions Center of Excellence 2. enable RCP at the project level and for every stage within the parallel shared container. Management. No part of this publication may be reproduced. stored in a retrieval system. whether the job or ETL processing is to continue is specified on a per-job and/or per-sequence and/or per-script basis based on business requirements. The presence of rejects and errors will be detected and notification sent by email to selected staff. . The Usage Analysis and Multi-Job Compile tools can be used to recompile jobs that use a shared container. all jobs that use a shared container must be recompiled when the container is changed.6 Parallel Shared Containers Parallel Shared Containers allow common logic to be shared across multiple jobs. 3. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. 2006 43 of 179 © 2006 IBM Information Integration Solutions. The presence of errors may not prevent further processing. or translated into any language in any form by any means without the written permission of IBM. and Production Automation. Both rejects and errors will be archived and placed in a special directory for evaluation or other action by support staff. These activities are the responsibility of job sequencers used to group jobs by some reasonable grain or by a federated scheduler. transcribed. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. Because Parallel Shared Containers are inserted when a job is compiled. All rights reserved. as well as standardized routines for use in Job Sequencers are discussed further in Parallel Framework Standard Practices: Administration. Error rows are those rows caused by unforeseen data events such as values too large for a column or text in an unsupported language. This allows the container input and output links to contain only the columns relevant to the container processing. userid and password. data rule violations or other reasons that would disqualify a row from processing. 3. For maximum component re-use.

3. Only records that match the given table definition and format are output. The Sequential File stage offers the following reject options: Option Description Continue Drop read failures from input stream. Rejects are tracked by count only. These actions include: 1. a shared container error handler can be used. 2 3 Rejects should not exist but should not stop the job. Parallel Framework Red Book: Data Flow Job Design July 17. A message is always written to the Director log which details the count of rows successfully read and rows rejected. Rejects should not exist and should stop the job.Information Integration Solutions Center of Excellence ETL actions to be taken for each record type is specified for each stage in the job design document. 2006 44 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. Use the Fail option.rej file extension. Reprocess – rows are reprocessed and re-enter the data stream. or translated into any language in any form by any means without the written permission of IBM. 3. Alternatively. (No reject link exists) Fail Abort job on read format failure (No reject link exists) Output Reject switch failures to the reject stream. it must have a *. Use the Output option. (Reject link exists) The reject option should be used in all cases where active management of the rejects is required. transcribed. If a file is created by this option.7. which outputs rows that do not match the given table definition and format specifications.1 Reject Handling with the Sequential File Stage The Sequential File stage can optionally include a reject link. Ignore – some process or event downstream of the ETL process is responsible for handling the error. stored in a retrieval system. . Send the reject stream to a *. The default action is to push back reject and error rows to a Data Steward. 2. All rights reserved. rows that cannot be read are dropped by the Sequential File stage. and should be reviewed by the Data Steward. By default.rej file. Pass successful reads to the output stream. Pass successful reads to the output stream. transmitted. Push back – rows are sent to a Data Steward for corrective action. Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Sequential File Stage Option Use the Continue option.

The reject option should be used in all cases where active management of the rejects is required.rej extension is used when rejects require investigation after a job run. All rights reserved. Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Rejects can exist in the data.Information Integration Solutions Center of Excellence 3. Drop Drop lookup failures from the input stream. Rejects should not exist but should not stop the job. and should be reviewed by the Data Steward. a local error handler based on a shared container can be used. No part of this publication may be reproduced. Fail Abort job on lookup failure Reject Reject lookup failures to the reject stream. Alternatively. Pass successful lookups to the output stream. and rejects can occur if the key fields are not found in the reference data. or translated into any language in any form by any means without the written permission of IBM. Pass successful lookups to the output stream.rej file or tag and merge with the output stream. 2 3 4 3.err extension when rejects can be ignored but need to be recorded.3 Reject Handling with the Transformer Stage Rejects occur when a transformer stage is used and a row: Parallel Framework Red Book: Data Flow Job Design July 17. to enforce error management ONLY ONE REFERENCE LINK is allowed on a Lookup stage. it must have a *. the *. Use the Fail option. . This behavior makes the Lookup stage very valuable for positive (reference is found) and negative (reference is NOT found) business rule validation. If there are multiple validations to perform. DS/EE offers the following options within a Lookup stage: Option Description Continue Ignore lookup failures and pass lookup fields as nulls to the output stream. however.rej or *. Send the reject stream to an *. stored in a retrieval system.err file extension.7. If a file is created by this option.7. each must be done in its own Lookup. Furthermore. The *. transmitted. Lookup Stage Option Drop if lookup fields are necessary down stream or Continue if lookup fields are optional Send the reject stream to an *.err file or tag and merge with the output stream. 2006 45 of 179 © 2006 IBM Information Integration Solutions. Pass successful lookups to the output stream. they only need to be recorded but not acted on. Rejects should not exist and should stop the job.2 Reject Handling with the Lookup Stage The Lookup stage compares a single input stream to one or more reference streams using keys. transcribed.

Rejects can exist in the data. Transformer Stage Option Funnel the reject stream back to the output stream(s). The *. Rejects are categorized in the ETL job design document using the following ranking: Category 1 2 3 4 Description Rejects are expected and can be ignored. however. . it must have a *.err file or tag and merge with the output stream. and Oracle Enterprise) offer an optional reject link that can be used to capture rows that cannot be written to the target database. and be reviewed by the Data Steward. Rejects should not exist but should not stop the job.4 Reject Handling with target database stages Some database stages (such as DB2/UDB Enterprise. 3.rej or *. Send the reject stream to an *. a shared container error handler can be used. Satisfies requirements for a reject conditional output stream.7. or translated into any language in any form by any means without the written permission of IBM. Otherwise. OR 2. 2006 46 of 179 © 2006 IBM Information Integration Solutions. they only need to be recorded but not acted on.Information Integration Solutions Center of Excellence 1.rej file extension. A message is always written to the Director log which details the count of rows successfully read and rows rejected. transcribed. If a file is created by this option.err extension when rejects can be ignored but need to be recorded. Send the reject stream to an *. Parallel Framework Red Book: Data Flow Job Design July 17. Cannot satisfy requirements of any conditional output stream and is rejected by the default output stream. The reject option should be used in all cases where active management of the rejects is required. Send the reject stream to a reject file and halt the job. Alternatively. a reject link must exist on that stage. To capture rejects from a target database. If a file is created from the reject stream. stored in a retrieval system. No part of this publication may be reproduced. transmitted. Target database stages offer the following reject options: Option No reject link exists Reject link exists Description Do not capture rows that fail to be written. a shared container error handler is used. reject rows will not be captured. Pass rows that fail to be written to the reject stream. the *. Alternatively. it must have a *.err file extension.rej file or tag and merge with the output stream. Rejects should not exist and should stop the job.rej extension is used when rejects require investigation after a job run. ODBC Enterprise. All rights reserved.

Rejects are tracked by count only.rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504. The job serial number (jpJOBSERIALNO) and a period “.Information Integration Solutions Center of Excellence Rejects are categorized in the ETL job design document using the following ranking: Category 1 Description Rejects are expected and can be ignored Target Database Stage Option No reject link exists.7. The standard columns for error and reject processing are: Column Name HOST_NAME Key? Yes Data Source DSHostName transformer macro in the error handler July 17. Only records that match the given table definition and database constraints are written.5 Error Processing Requirements Jobs will produce flat files containing reject and errors and may alternatively process rows on reject ports and merge these rows with the normal output stream. transcribed. transmitted. or translated into any language in any form by any means without the written permission of IBM.”. and 5.rej file.5. For example.err Rows will be converted to the common file record format with 9 columns (below) using Column Export and Transformer stages for each reject port. Reject link exists. and should be reviewed by the Data Steward. The project phase (jpENVIRON) and a underscore “_”. The job name (jpJOB_NAME) and a underscore “_”. and gathered using a Funnel stage that feeds a Sequential File stage. The project name (jpPROJECT_NAME) and a underscore “_”. All rights reserved. Send the reject stream to a *. The appropriate file type. No part of this publication may be reproduced. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. 3. This section deals with both methods of handling errors.1 Processing Errors and Rejects to a Flat File Each job will produce a flat file for errors and a flat file for rejects with a specific naming convention: 1. 2. stored in a retrieval system. The Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job. 3. 2 Rejects should not exist but should not stop the job. . 4. 3.7. one of “rej” or “err”. job DECRP_N_XformClients in the ECR_FACTS project in the development environment with a serial number of 20060201-ETL-091504 would have these reject and error file names: ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504.

The Transformer stage adds the required key columns. or translated into any language in any form by any means without the written permission of IBM. the following stages process the only errors produced by a job: The Column Export stage maps the unique columns to the single standard column. stored in a retrieval system. No part of this publication may be reproduced. transcribed. All rights reserved. transmitted.Information Integration Solutions Center of Excellence PROJECT_NAME JOB_NAME STAGE_NAME DATA_OBJ_NAME JOB_SERIALNO ETL_ROW_NUM ETL_BAT_ID ROW_DATA Yes Yes Yes Yes Yes Yes Yes No DSProjectName transformer macro in the error handler DSJobName transformer macro in the error handler The name of the stage from which the error came The source table or file data object name jpJOBSERIALNO Data stream coming in to the error handler Data stream coming in to the error handler The columns from the upstream stages reject port exported to a single pipe-delimited “|” varchar(2000) column using the Column Export stage in the error handler In this example. . 2006 48 of 179 © 2006 IBM Information Integration Solutions. Figure 13: Error Processing Components Parallel Framework Red Book: Data Flow Job Design July 17.

Information Integration Solutions Center of Excellence The input to the Column Export stage explicitly converts the data unique to the reject stream (in this case. Track*) to a single output column. All rights reserved. . ROW_DATA: Figure 14: Error Processing Column Export stage And the downstream Transformer stage builds the standard output record by creating the required keys: Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. or translated into any language in any form by any means without the written permission of IBM. 2006 49 of 179 © 2006 IBM Information Integration Solutions. transcribed. stored in a retrieval system. No part of this publication may be reproduced.

OR with columns contain illegal values for some operation performed on said columns.2 Processing Errors and Rejects and Merging with an Output Stream There may be processing requirements that specify that rejected or error rows be tagged as having failed a validation and merged back into the output stream. Connect the reject port to a Transformer stage where columns are set to specific values. 2006 50 of 179 © 2006 IBM Information Integration Solutions. . Connect the output stream of the Transformer stage and one or more output streams of the Switch stage to a Funnel stage to merge the two (or more) streams. This is done by processing the rows from the reject ports and setting the value of a specific column with a value specified by the design document. A failed switch will reject an intact input row show key fails to resolve to one of the Switch output stream. rows rejected by the Lookup stage are processed by a corrective Transformer stage where the failed references as set to a specific value and then merged with the output of the Lookup stage: Parallel Framework Red Book: Data Flow Job Design July 17. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. In either case. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key. stored in a retrieval system. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream. or translated into any language in any form by any means without the written permission of IBM. Switch Transformer In this example. Connect the output stream of the corrective Transformer stage and one or more output streams of the original Transformer stage to a Funnel stage to merge the two (or more) streams.5. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values. The following table identifies the tagging method to be used for the previously cited operators. transcribed. All rights reserved. Connect the reject port to a Transformer stage where columns are set to specific values. One or more columns may have been selected for replacement when a reference key is found.7. Connect the output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams. transmitted.Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3. No part of this publication may be reproduced.

transmitted.1 Server Edition Components Avoid the use of Server Edition components in parallel job flows. 3.8. This section provides guidelines appropriate use of various stages when building a parallel job flows.8 Component Usage DataStage Enterprise Edition offers a wealth of component types for building ETL flows. stored in a retrieval system. 2006 51 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Figure 16: Error Processing Lookup example 3. No part of this publication may be reproduced. The ability to use a Server Edition component within a parallel job is intended only as a migration option for existing Server Edition applications that might benefit by leveraging some parallel capabilities on SMP platforms. transcribed. severely impacting scalability and network resources. In clustered an MPP environments Server Edition components only run on the primary (conductor) node. Server Edition components limit overall performance of large-volume job flows since many components such as the BASIC Transformer use interpreted psuedo-code. .

Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: .8. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime. . 3. it can be used at the end of a data flow o For simple jobs with only two stages. and in Enterprise Edition native format. and Production Automation.BASIC Routines . 3. 3. For this reason. and necessary. it is best to develop a job iteratively using the Copy stage as a “placeholder”.3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs.8. NOTE: Because parallel Data Sets are platform and configuration-specific.Server shared containers Note that BASIC Routines are still appropriate. the Copy stage should be used as a placeholder so that new stages can be inserted easily should future requirements change.4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. Used in this manner. This is discussed in more detail in Parallel Framework Standard Practices: Administration. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run.Renaming Columns .BASIC Transformers .Job Design placeholder between stages (unless the Force option =true. Enterprise Edition will optimize this out at runtime) . Since the Copy stage does not require an output link. they should not be used for long-term archive of source data. stored in a retrieval system. • The Copy stage should be used instead of a Transformer for simple operations including: . always write to parallel Data Sets. transcribed.2 Copy Stage For complex data flows. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs. All rights reserved. and to use other stages (such as Copy) when derivations are not needed. o Unless the Force property is set to “True”.Dropping Columns July 17. Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form. it is important to minimize the number of transformers. transmitted. 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Management. in sort order. Data Sets offer parallel I/O on read and write operations. without overhead for format or data type conversions.8.

transmitted. All rights reserved.2. Parallel Framework Red Book: Data Flow Job Design July 17. implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible. 2006 53 of 179 © 2006 IBM Information Integration Solutions. Consider. No part of this publication may be reproduced. .1.Complex reusable logic cannot be implemented using the Transformer or .8.5 BuildOp stages BuildOps should only be used when: . if possible. user-defined functions and routines can expand parallel Transformer capabilities. See Section 8.2: Modify Stage.1.Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4. The only time that Filter or Switch should be used is when the selection clauses need to be parameterized at runtime. Instead. and default type conversion can also be performed by the output mapping tab of any stage.5. • Because the parallel Transformer is compiled. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines.7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B.3 Then B=”X” If A=4. null handling.6. For example.Existing Transformers do not meet performance requirements As always. it is faster than the interpreted Filter and Switch stages. drop (if Runtime Column Propagation is disabled). 3. or translated into any language in any form by any means without the written permission of IBM. Even then its use should be restricted and the routines should be converted as soon as possible. • NEVER use the “BASIC Transformer” stage in large-volume job flows. transcribed.2: Default and Explicit Type Conversions] Note that rename. and character string trimming. the derivation expression: If A=0. stored in a retrieval system. • The Modify stage can be used for non-default type conversions. performance should this should be tested in isolation to identify specific cause of bottlenecks.

or translated into any language in any form by any means without the written permission of IBM. consisting of a fixed or variable number of contiguous bytes and an optional alignment value ASCII character string of fixed or variable length (Unicode Extended option NOT selected) ASCII character string of fixed or variable length (Unicode Extended option NOT selected) ASCII character string of fixed or variable length (Unicode Extended option IS selected) Time of day. unit32 int64. VarBinary Unknown. day. LongVarChar. (Specify microseconds Extended option) string 1 byte per character ustring ustring time time(microseconds) timestamp timestamp(microseconds) multiple bytes per character multiple bytes per character 5 bytes 5 bytes 9 bytes 9 bytes 1 BigInt values map to long long integers on all supported platforms except Tru64 where they map to longer integer values. All rights reserved. Real Double TinyInt SmallInt Integer BigInt1 Binary. The following table summarizes the underlying data types of DataStage Enterprise Edition: SQL Type Internal Type Size Description Date Decimal. transcribed. VarChar Time Time Timestamp Timestamp date decimal sfloat dfloat int8. LongNVarChar Char. VarChar NChar. No part of this publication may be reproduced. IEEE single-precision (32-bit) floating point value IEEE double-precision (64-bit) floating point value Signed or unsigned integer of 8 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 16 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 32 bits (Specify unsigned Extended option for unsigned) Signed or unsigned integer of 64 bits (Specify unsigned Extended option for unsigned) Untyped collection. Each SQL data type maps to an underlying data type in the Enterprise Edition engine. NVarChar. 2006 54 of 179 © 2006 IBM Information Integration Solutions. LongVarChar. transmitted. Single field containing both date and time value with resolution to microseconds. Parallel Framework Red Book: Data Flow Job Design July 17. unit64 raw 4 bytes (roundup(p)+1)/2 4 bytes 8 bytes 1 byte 2 bytes 4 bytes 8 bytes 1 byte per character Date with month.Information Integration Solutions Center of Excellence 4 DataStage Data Types The DataStage Designer and Manager represent column data types using SQL notation. Numeric Float. uint8 int16. with resolution to seconds Time of day. LongVarBinary. . and year Packed decimal. stored in a retrieval system. with resolution of microseconds (Specify microseconds Extended option) Single field containing both date and time value with resolution to seconds. uint16 int32. The internal Enterprise Edition data types are used in schema files and are displayed when viewing generated OSH or viewing the output from $OSH_PRINT_SCHEMAS. Char. compatible with IBM packed decimal format. Bit.

and LongVarChar SQL types relate to underlying string types where each character is 8-bits and does not require mapping because it represents an ASCII character. . transcribed. NVarChar.1. VarChar. ustring data represents full Unicode (UTF-16) data. Default type conversions take place across the stage output mappings of any Enterprise Edition stage. VarChar.1. specify that these data types are extended.) An Extended field appears in the columns grid.2 Default and Explicit Type Conversions DataStage Enterprise Edition provides a number of default conversions and conversion functions when mapping from a source to a target data type. parallel jobs support two types of underlying character data types: strings and ustrings. 4. The following table summarizes Data Type conversions: Source Field Target Field d = There is a default type conversion from source field type to destination field type.Information Integration Solutions Center of Excellence 4. or translated into any language in any form by any means without the written permission of IBM. The NChar. and extended Char. stored in a retrieval system. e = You can use a Modify or a Transformer conversion function to explicitly convert from the source field type to the destination field type. (They are specified as such by selecting the Extended check box for the column in the Edit Meta Data dialog box. String data represents unmapped bytes. No part of this publication may be reproduced.1 Strings and Ustrings If NLS is enabled on your DataStage server. The Char. All rights reserved. or LongVarChar columns have ‘Unicode’ in this field. You can. date uint8 sfloat uint16 uint32 uint64 dfloat int16 int32 int64 string ustring time e e e int8 timestamp e e e e 55 of 179 decimal raw Int8 uint8 Int16 uint16 Int32 uint32 Int64 uint64 sfloat dfloat decimal string d de d de d de d de de de de d d d d d d d d d d d d d d d d d d d d d de d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d d de d d d d d d d d de de d d d d d d d d d d e d d d d d d d d d d d d d d de d d d d d d de d d d de d d d de d d e d d e d e d e d e d d d d e d e de d de de de de d d d de de d e e e e Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. in which case they are taken as ustrings and do require mapping. however. A blank cell indicates that no conversion is provided. and LongNVarChar types relate to underlying ustring types so do not need to be explicitly extended. 2006 © 2006 IBM Information Integration Solutions.

2006 56 of 179 © 2006 IBM Information Integration Solutions. . • 4. Sequential File and DB2/UDB Enterprise targets) allow the pad character to be specified in their stage or column definition properties. Enterprise Edition pads the remaining length with NULL (ASCII zero) characters by default. depending on the source and result data types.It designates a specific field value to indicate a null. for example a numeric field’s most negative possible value. an ASCII space (0x20) or a Unicode space (U+0020). Enterprise Edition displays a warning message in the job log.2 Null Handling DataStage Enterprise Edition represents nulls in two ways: . for example. When entering a space for the value of APT_STRING_PADCHAR do note enclose the space character in quotes. You must first convert a Char string type to a Varchar type before using PadString. All rights reserved. • The environment variable APT_STRING_PADCHAR can be used to change the default pad character from an ASCII NULL (0x0) to another character. transcribed. This type of representation is called an out-of-band null. Parallel Framework Red Book: Data Flow Job Design July 17. Note that PadString does not work with fixed-length (CHAR) string types.It allocates a single bit to mark a field as null. Some stages (for example. When converting from variable-length to fixed-length strings. the specified pad character will override the default for that stage only. In-band null representation can be disadvantageous because you must reserve a field value for nulls and this value cannot be treated as valid data elsewhere. • As an alternate solution. or translated into any language in any form by any means without the written permission of IBM. . transmitted. When used in these stages.Information Integration Solutions Center of Excellence ustring raw date time timesta mp de e e e e d de e d d e e e e de e d d d de e d e d e e e e e e e e e de e e e e The conversion of numeric data types may result in a loss of precision and cause incorrect results. No part of this publication may be reproduced. the PadString Transformer function can be used to pad a variablelength (Varchar) string to a specified length using a specified pad character. The Transformer and Modify stages can change a null representation from an out-of-band null to an inband null and from an in-band null to an out-of-band null. This type of representation is called an in-band null. stored in a retrieval system. In these instances.

When reading from or writing to Sequential Files or File Sets. out-of-band null representation for NULL values. transcribed.Information Integration Solutions Center of Excellence IMPORTANT: When processing nullable columns in a Transformer stage. . If the source value is not null. All rights reserved. See Section 8. Source value propagates. or translated into any language in any form by any means without the written permission of IBM. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. as shown in Figure 17: Figure 17: Extended Column Metadata (Nullable properties) The Table Definition of a stage’s input or output data set can contain columns defined to support outof-band nulls (Nullable attribute is checked). destination value is never null. the source value propagates. care must be taken to avoid data rejects. When reading from Data Set and database sources with nullable columns. No part of this publication may be reproduced. 2006 57 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Source value or null propagates. The next table lists the rules for handling nullable fields when a stage takes a Data Set as input or writes to a Data Set as output. stored in a retrieval system. Enterprise Edition uses the internal. Source Field not Nullable Nullable not Nullable Nullable Destination Field not Nullable Nullable Nullable not Nullable Result Source value propagates to destination. transmitted.1.1: Transformer NULL Handling and Reject Link. July 17.

4. . RCP facilitates re-use through parallel shared containers. Furthermore. the remaining columns pass through at runtime. Parallel Framework Red Book: Data Flow Job Design July 17. allow their runtime schema to be parameterized further extending re-use through RCP. transmitted. rather than using a large number of jobs with hard-coded table definitions to perform the same tasks. Before a DataStage developer can use RCP. Using RCP judiciously in a job design facilitates re-usable job designs based on input metadata. for example the Sequential File stage. 2006 58 of 179 © 2006 IBM Information Integration Solutions. Some stages. No part of this publication may be reproduced. as long as each stage in the shared container has RCP enabled on their stage Output properties.Information Integration Solutions Center of Excellence If the source value is null. it must be enabled at the project level through the Administrator client. All rights reserved. transcribed.3 Runtime Column Propagation Runtime column propagation (“RCP”) allows job designs to accommodate additional columns beyond those defined by the job developer. only the columns explicitly referenced within the shared container logic need to be defined. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. a fatal error occurs. Using RCP.

Remove Duplicates) or stage requirements (for example. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. a “fan-out” icon is drawn on the link within the Designer canvas. No part of this publication may be reproduced. different types of partitioning are provided for the parallel job developer. Partitioners exist before any stage that is running in parallel. and are indicated by a “fan-in” icon as shown in this example: Stage running in parallel Figure 19: Collector icon This section provides an overview of partitioning and collecting methods. ensuring that records with the same values in those key column(s) are assigned to the same partition. it is important that this distribution does not violate business requirements for accurate data processing. as shown in this example: Stage running sequentially Figure 18: “fan-out” icon Collectors combine parallel partitions of a single link for sequential processing. and provides guidelines for appropriate use in job designs.Information Integration Solutions Center of Excellence 5 Partitioning and Collecting Partition parallelism is a key to establishing scalable performance of DataStage Enterprise Edition. or translated into any language in any form by any means without the written permission of IBM. Collectors only exist before stages running sequentially and when the previous stage is running in parallel. Stage running sequentially Stage running in parallel 5. If the prior stage was running sequentially. For this reason. It also provides tips for monitoring jobs running in parallel. transmitted. 2006 59 of 179 © 2006 IBM Information Integration Solutions. Keyed partitioning examines the data values in one or more key columns.Keyless partitioning distributes rows without regard to the actual data values. Keyed partitioning is used when business rules (for example.1 Partition Types While partitioning allows data to be distributed across multiple processes running in parallel. Partitioners distribute rows of a single link into smaller segments that can be processed independently in parallel. Join) require processing on groups of related records. Partitioning methods are separated into keyless and keyed classes: . . Different types of keyless partitioning methods define the method of data distribution. stored in a retrieval system.

2006 60 of 179 © 2006 IBM Information Integration Solutions. Furthermore. The partitioning method is specified in the Input stage properties using the “Partitioning” option as shown on the right: Figure 20: Specifying Partition method 5. Auto partitioning specifies that the Enterprise Edition engine will attempt to select the appropriate partitioning method at runtime. Within the Designer canvas. However. .Information Integration Solutions Center of Excellence The default partitioning method for newly-drawn links is Auto partitioning. The “Preserve Partitioning” flag is an internal “hint” that Auto partitioning uses to attempt to preserve carefully ordered data (for example.1. the partitioning method may not necessarily be the most efficient from an overall job perspective. links with Auto partitioning are drawn with the following link icon: Figure 21: Auto partitioning icon Auto partitioning is designed to allow the beginning DataStage developer to construct simple data flows without having to understand the details of parallel design principles. to improve performance. However. This flag is set automatically by Parallel Framework Red Book: Data Flow Job Design July 17.1 Auto Partitioning The default partitioning method for newly-drawn links. Entire) and keyed (Hash) partitioning methods to produce functionally correct results and. or translated into any language in any form by any means without the written permission of IBM. Round Robin. Data Sets. and job design (stage requirements and properties). transmitted. All rights reserved. the ability for the Enterprise Edition engine to determine the appropriate partitioning method depends on the information available to it. on the output of a parallel Sort). if the logic defined in a Transformer stage is based on a group of related records. Based on the configuration file. For example. since the Enterprise Edition engine has no visibility into user-specified logic (such as Transformer or BuildOp stages) it may be necessary to explicitly specify a partitioning method for some stages. Auto partitioning will ensure correct results when using built-in stages. Auto partitioning will select between keyless (Same. in some cases. then a keyed partitioning method must be specified to achieve correct results. transcribed. No part of this publication may be reproduced. stored in a retrieval system. In general.

No part of this publication may be reproduced. as shown in on the right: Figure 22: Preserve Partitioning option The Preserve Partitioning flag is part of the Data Set structure.1 Same Partitioning Same partitioning in fact performs no partitioning to the input Data Set. In these instances. transcribed. it retains the partitioning from the output of the upstream stage. between servers).1. 5. . There are some cases when the input stage requirements prevent partitioning from being preserved. Each partition receives the entire Data Set.1. if the Preserve Partitioning flag was set. 5. as illustrated on the right: Same partitioning doesn’t move data between partitions (or. Distributes rows evenly across partitions in a random partition assignment. for example). the degree of parallelism remains unchanged: Parallel Framework Red Book: Data Flow Job Design July 17. and is appropriate when trying to preserve the grouping of a previous operation (for example. or translated into any language in any form by any means without the written permission of IBM.2 Keyless Partitioning Keyless partitioning methods distribute rows without examining the contents of the data: Keyless Partition Method Same Round Robin Random Entire Description Retains existing partitioning from previous stage. Distributes rows evenly across partitions. and its state is stored in persistent Data Sets. a parallel Sort). Instead. transmitted. in a round robin partition assignment. Because Same does not redistribute existing partitions. 2006 61 of 179 © 2006 IBM Information Integration Solutions. although it can be explicitly set or cleared in the “Advanced” stage properties of a given stage. a warning will be placed in the Director log indicating that Enterprise Edition was unable to preserve partitioning for a specified stage. All rights reserved.2. links that have been specified with Same partitioning are drawn with a “horizontal line” partitioning icon: Figure 23: Same partitioning icon It is important to understand the impact of Same partitioning in a given data flow. in the case of a cluster or Grid.Information Integration Solutions Center of Excellence some stages (Sort. Row ID's 0 3 6 1 4 7 2 5 8 0 3 6 1 4 7 2 5 8 Within the Designer canvas. stored in a retrieval system.

4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition. . as the complete Data Set must be distributed across the network to each node. 2006 62 of 179 © 2006 IBM Information Integration Solutions.1. On clustered and Grid implementations.2.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially. or translated into any language in any form by any means without the written permission of IBM. Entire partitioning may have a performance impact. Random partitioning has a slightly higher overhead than Round Robin partitioning. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17. Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning. transmitted. Since the random partition number must be calculated. it is rarely used in real-world data flows. as illustrated on right: Entire partitioning is useful for distributing the reference data of a Lookup task (this may or may not involve the Lookup stage). No part of this publication may be reproduced. 5. the order that rows are assigned to a particular partition will differ between job runs. 3 2 1 0 . transcribed. Random partitioning evenly distributes rows across partitions. While in theory Random partitioning is not subject to regular data patterns that might exist in the source data. .2.2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment. the downstream stage runs with the degree of parallelism used to create the Data Set. All rights reserved. but using a random assignment. regardless of the current $APT_CONFIG_FILE 5. Since optimal parallel processing occurs when all partitions have the same workload. . 3 2 1 0 .2. .1. …8 7 6 5 4 3 2 1 0 ENTIRE . stored in a retrieval system. As a result.1.3 Random Partitioning Like Round Robin. similar to dealing cards: Round robin partitioning has a fairly low overhead.

3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns.1. No part of this publication may be reproduced. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. For DB2 Enterprise Server Edition with DPF (DB2/UDB) only – matches the internal partitioning of the specified source or target table. If the source data values are evenly distributed within these key column(s).1. Remove Duplicates) or stage requirements (for example. Keyed partitioning is used when business rules (for example. 2006 63 of 179 © 2006 IBM Information Integration Solutions. 5. consider the following sample Data Set: ID 1 2 3 4 5 6 7 8 9 10 LName Ford Ford Ford Ford Dodge Dodge Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Horace John Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 17840 Jefferson 75 Boston Boulevard 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Values of key column …0 3 2 1 0 2 3 2 1 1 HASH 0 3 0 3 1 1 1 2 2 2 Parallel Framework Red Book: Data Flow Job Design July 17. As an example of hashing. stored in a retrieval system.Information Integration Solutions Center of Excellence 5.3. transmitted. then the resulting partitions will be of relatively equal size. . ensuring that records with the same values in those key column(s) are assigned to the same partition.1 Hash Partitioning Hash partitioning assigns rows with the same values in one or more key column(s) to the same partition using an internal hashing algorithm. and there are a large number of unique values. Keyed Partition Method Hash Modulus Range DB2 Description Assigns rows with the same values in one or more key column(s) to the same partition using an internal hashing algorithm. All rights reserved. Assigns rows with the same values in one or more key column(s) to the same partition using a specified range map generated by pre-reading the Data Set. Join) require processing on groups of related records. or translated into any language in any form by any means without the written permission of IBM. transcribed.

transcribed. No part of this publication may be reproduced. stored in a retrieval system. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. . transmitted. 2006 64 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence Hashing on key column LName would produce the following results: Partition 0: ID 5 6 LName Dodge Dodge FName Horace John Address 17840 Jefferson 75 Boston Boulevard Partition 1: ID 1 2 3 4 7 8 9 10 LName Ford Ford Ford Ford Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore Parallel Framework Red Book: Data Flow Job Design July 17.

which would impact performance. the partition size of modulus partitioning will be equally distributed as long as the data values in the key column are equally distributed.2 Modulus Partitioning Modulus partitioning uses a simplified algorithm for assigning related records based on a single integer key column. hash partitioning on the key columns LName and FName yields the following distribution with a 4-node configuration file: Partition 0: I D 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen I D 4 6 10 LName Ford Dodge Ford Partition 2: FName Eleanor John Eleanor Address 7900 Jefferson 75 Boston Boulevard 1100 Lakeshore Partition 1: I D 3 5 9 LName Ford Dodge Ford FName Edsel Horace Edsel Address 7900 Jefferson 17840 Jefferson 1100 Lakeshore I D 1 7 LName Ford Ford Partition 3: FName Henry Henry Address 66 Edison Avenue 4901 Evergreen In this example. The remainder is used to assign the value to a given partition: partition = MOD (key_value / number of partitions) Like hash. stored in a retrieval system. . individual key column values have no significance for partition assignment. transmitted.1. producing partition skew. 2006 65 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. there are more instances of “Ford” than “Dodge”. regardless of the actual number of nodes in the parallel configuration file. Using the same source Data Set. or translated into any language in any form by any means without the written permission of IBM. the key column combination of LName and FName yields improved data distribution and a greater degree of parallelism. Parallel Framework Red Book: Data Flow Job Design July 17. When using hash partitioning on a composite key (more than one key column).3.Information Integration Solutions Center of Excellence In this case. Also note that in this example the number of unique values will limit the degree of parallelism. transcribed. Also note that only the unique combination of key column values appear in the same partition when used for hash partitioning. All rights reserved. It performs a modulus operation on the data value using the number of partitions as the divisor. 5.

3. transcribed.Information Integration Solutions Center of Excellence Since modulus partitioning is simpler and faster than hash. Modulus partitioning cannot be used for composite keys. these rows will be assigned to either the first or the last partition. All rights reserved. Range partitioning must read the Data Set twice: once to create a Range Map file.3.1. depending on the value. Range partitioning will result in sequential processing.4 DB2 Partitioning The DB2/UDB Enterprise stage matches the internal database partitioning of the source or target DB2 Enterprise Server Edition with Data Partitioning Facility database (previously called “DB2/UDB EEE”).1. And. the Range Map file can be re-used. transmitted. . use Same partitioning on the input to downstream stages. In these instances. it should be used if you have a single integer key column. Using the DB2/UDB Enterprise stage. typically where the incoming data values and distribution are consistent over time. 2006 66 of 179 © 2006 IBM Information Integration Solutions. if new data values are processed outside of the range of a given Range Map. To maintain partitioning on data read from a DB2/UDB Enterprise stage. A Range Map file is specific to a given parallel configuration file. or for a non-integer key column. defeating the intention of Range partitioning. 5. To achieve this balanced distribution. partition balance will be skewed. Also. unlike Hash and Modulus partitioning where partition skew is dependent on the actual data distribution. or translated into any language in any form by any means without the written permission of IBM. 5. and the second to actually partition the data within a flow using the Range Map. stored in a retrieval system.3 Range Partitioning As a keyed partitioning method. Range partitioning assigns rows with the same values in one or more key column(s) to the same partition. data is read in parallel from each DB2 node. In another scenario to avoid. if the incoming Data Set is sequential and ordered on the key column(s). by default. Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. DB2 partitioning can only be specified for target DB2/UDB Enterprise stages. Given a sufficient number of unique values. Values of key column 4 0 5 1 6 0 5 4 3 RANGE Rang e Map fi le 0 1 0 4 4 3 The “read twice” penalty of Range partitioning limits its use to specific scenarios. Range partitioning ensures balanced workload by assigning an approximately equal number of rows to each partition. data is partitioned to match the internal partitioning of the target DB2 table using the DB2 partitioning method. when writing data to a target DB2 database using the DB2/UDB Enterprise stage. It is important to note that if the data distribution changes without recreating the Range Map.

DataStage Enterprise Edition determines the degree of parallelism for each stage using: a) the parallel configuration file (APT_CONFIG_FILE) b) the degree of parallelism of existing source and target Data Sets (and. This is very useful in determining the distribution across parallel partitions (skew). The partition number is appended to the stage name for messages written to the Director log. All rights reserved. stored in a retrieval system. Partitions are assigned numbers. transcribed.Information Integration Solutions Center of Excellence 5. as shown in the example log below where the stage named “Peek” is running with four degrees of parallelism (partition numbers zero through 3): Figure 24: Partition numbers as shown in Director log To display row counts per partition in the Director Job Monitor window. and each stage is processing an equal number (12. as illustrated below. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. in some cases. starting at zero. or translated into any language in any form by any means without the written permission of IBM.4. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. . 2006 67 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.2 Monitoring Partitions At runtime. transmitted. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True. databases) c) and. right-click anywhere in the window. and select the “Show Instances” option.500) of rows for an optimal balanced workload. In this instance. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name).2Understanding the Parallel Job Score. if specified. Figure 25: Director Job Monitor row counts by partition Setting the environment variable APT_RECORD_COUNTS will output the row count per link per partition to the Director log as each stage/node completes processing. Specific details on interpreting the parallel job score can be found in 12.

transmitted. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. the “Data Set Management” tool (available in the Tools menu of Designer. 5.Information Integration Solutions Center of Excellence Figure 26: Output of APT_RECORD_COUNTS in Director log Finally. transcribed. . as shown below: Figure 27: Data Set Management Tool In a non-graphical way. while minimizing overhead. All rights reserved. 2006 68 of 179 © 2006 IBM Information Integration Solutions. the following objectives help to form a methodology for assigning partitioning: Objective 1: Choose a partitioning method that gives close to an equal number of rows in each partition. stored in a retrieval system.3 Partition Methodology Given the numerous options for keyless and keyed partitioning. the orchadmin command line utility on the DataStage server can also be used to examine a given parallel Data Set. Director. This ensures that the processing workload is evenly balanced. minimizing overall run time. Parallel Framework Red Book: Data Flow Job Design July 17. or Manager) can be used to identify the degree of parallelism and number of rows per partition for an existing persistent Data Set.

This is particularly useful if downstream jobs are run with the same degree of parallelism (configuration file) and require the same partition and sort order. transmitted. Change Capture. All rights reserved. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. Objective 4: Partition method should not be overly complex The simplest method that meets the above objectives will generally be the most efficient and yield the best performance. Note that in satisfying the requirements of this second objective. persistent Data Sets can be used to retain the partitioning and sort order. It may also be necessary for Transformers and BuildOps that process groups of related records. o Across jobs. assigning related records to the same partition if required Any stage that processes groups of related records (generally using one or more key columns) must be partitioned using a keyed partition method. . Using the above objectives as a guide. No part of this publication may be reproduced. 2006 69 of 179 © 2006 IBM Information Integration Solutions. minimize repartitioning. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport.Information Integration Solutions Center of Excellence Objective 2: The partition method must match the business requirements and stage functional requirements. Change Apply. Merge. Join. Remove Duplicates. and Sort stages. Parallel Framework Red Book: Data Flow Job Design July 17. This includes. transcribed. the following methodology can be applied: a) Start with Auto partitioning (the default) b) Specify Hash partitioning for stages that require groups of related records o Specify only the key column(s) that are necessary for correct grouping as long as the number of unique values is sufficient o Use Modulus partitioning if the grouping is on a single integer key column o Use Range partitioning if the data is highly skewed and the key column values and distribution do not change significantly over time (Range Map can be reused) c) If grouping is not required. but is not limited to: Aggregator. use Round Robin partitioning to redistribute data equally across all partitions o Especially useful if the input Data Set is highly skewed or sequential d) Use Same partitioning to optimize end-to-end partitioning and to minimize repartitioning o Being mindful that Same partitioning retains the degree of parallelism of the upstream stage o Within a flow. Objective 3: Unless partition distribution is highly skewed. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit).

on closer inspection. a new output column is defined on the header and detail links using a single constant value derivation.4.Information Integration Solutions Center of Excellence 5. The output of the Aggregator is then sent to the second input of the Join.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. or through Auto partitioning): Parallel Framework Red Book: Data Flow Job Design July 17. . The standard solution would be to Hash partition (and Sort) the inputs to the Join and Aggregator stages as shown below: Figure 28: “Standard” Partitioning assignment However.4 Partitioning Examples In this section. Additional. stored in a retrieval system. we’ll apply the partitioning methodology defined earlier to several example job flows. To add aggregate columns to every detail row.4. transcribed. All rights reserved. more advanced partitioning and sorting examples are given in 12. No part of this publication may be reproduced.2 Partitioning Example 2 – Use of Entire Partitioning In this example.4. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator. Using a “standard” solution. 5. the partitioning and sorting of this scenario can be optimized. as shown below: Figure 29: Optimized Partitioning assignment This example will be revisited in the Sorting discussion because there is one final step necessary to optimize the sorting in this example. a Transformer is used to extract data from a single header row of an input file. we can move the Hash partition and Sort before the Copy stage. Because the Join and Aggregator use the same partition keys and sort order. Within the Transformer. or translated into any language in any form by any means without the written permission of IBM. and apply Same partitioning to the downstream links. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly.2Sorting and Hashing Advanced Example. 2006 70 of 179 © 2006 IBM Information Integration Solutions. transmitted. 5. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row.

Remembering that the degree of parallel operation is limited by the number of distinct values. In order to process a large number of detail records.Use Entire partitioning on the header input to copy the single header row to all partitions Figure 31: Optimized Partitioning assignment based on business requirements Because we are joining on a single value. stored in a retrieval system. transcribed. and the detail rows are assigned to the Left input as shown in the following illustration: Figure 32: Specifying Link Order in Join stage Parallel Framework Red Book: Data Flow Job Design July 17. the single value join column will assign all rows to a single partition. An optimized solution would be to alter the partitioning for the input links to the Join stage: . it is not always the most efficient solution. . No part of this publication may be reproduced.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions . consider that the single header row is really a form of reference data. To optimize partitioning. 2006 71 of 179 © 2006 IBM Information Integration Solutions. the link order of the Inner Join is significant. All rights reserved. For this reason. the link order in this example should be set so that the single header row is assigned to the Right input. the above solution has one serious limitation. there is no need to pre-sort the input to the Join. Although functionally correct. so we will revisit this in the Sorting discussion. depending on the business requirements. transmitted. The Join stage operates by reading a single row from the Left input and reading all rows from the Right input that match the key value(s).Information Integration Solutions Center of Excellence Figure 30: “Standard” Partitioning assignment for a Join stage Although Hash partitioning guarantees correct results for stages that require groupings of related records. resulting in sequential processing. or translated into any language in any form by any means without the written permission of IBM.

5.5. Consider an example where data is read sequentially and passed to a Round Robin partitioner: Parallel Framework Red Book: Data Flow Job Design July 17. Like partitioning methods. or translated into any language in any form by any means without the written permission of IBM. the order of rows in an Auto collector is undefined. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5. 2006 72 of 179 © 2006 IBM Information Integration Solutions. For advanced users. Because the Join will wait until it receives an End of Group (new key value) or End of Data (no more rows on the input Data Set) from the Right input. transcribed. and may vary between job runs on the same Data Set. the detail rows in the Left input will buffer to disk to prevent a deadlock. the collector method is defined in the stage Input/Partitioning properties for any stage running sequentially. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition. 5. 5.Information Integration Solutions Center of Excellence If defined in reverse of this order.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially. Changing the output derivation on the header row to a series of numbers instead of a constant value will establish the End of Group and prevent buffering to disk. All rights reserved. However.3: Minimizing Runtime Processes and Resource Requirements). For this reason. Auto is the default collector method. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory.2 Round Robin Collector The Round Robin collector patiently reads rows from partitions in the input Data Set by reading input partitions in round robin order. stored in a retrieval system. No part of this publication may be reproduced. . transmitted. there is one further detail in this example. (See Section 12. there is a specialized example where the Round Robin collector may be appropriate.1 Auto Collector The Auto collector eagerly reads rows from partitions in the input Data Set without blocking if a row is unavailable on a particular partition.

All rights reserved. transcribed. as used to sort the input Data Set in parallel. then a Round Robin collector can be used before the final Sequential output to reconstruct a sequential output stream in the same order as the input data stream. This is because Round Robin collector reads from partitions using the same partition order that a Round Robin partitioner assigns rows to parallel partitions. use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. 2006 73 of 179 © 2006 IBM Information Integration Solutions. 5. then reads all rows from the next partition until all rows in the Data Set have been collected. and these should be the same columns. stored in a retrieval system. The Sort Merge collector requires one or more key column(s) to be defined. In this scenario. the Ordered collector may be more efficient c) Use a Round Robin collector to reconstruct rows in input order for round-robin partitioned input Data Sets. .6 Collecting Methodology Given the options for collecting data into a sequential stream. use Sort Merge collector to produce a single. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter.5. in the same order. through aggregation). No part of this publication may be reproduced. as long as the Data Set has not been repartitioned or reduced. an Ordered collector will generate a sequential stream in sort order. transmitted. or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. Row order is undefined for non-key columns.4 Sort Merge Collector If the input Data Set is sorted in parallel.Information Integration Solutions Center of Excellence Round Robin partitioner Round Robin collector Sequential input Stage running in parallel Sequential output Figure 34: RoundRobin Collector example Assuming the data is not repartitioned within the job flow and that the number of rows is not reduced (for example. Ordered collectors are generally only useful if the input Data Set has been Sorted and Range partitioned on the same key column(s).5.3 Ordered Collector An Ordered collector reads all rows from the first partition. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. 5. the Sort Merge collector will generate a sequential stream of rows in globally sorted order. 5.

Change Apply. stored in a retrieval system. Join.1 Partition and Sort Keys Using the parallel Sort within DataStage Enterprise Edition: . This sort can be done in parallel. transcribed. 6. optionally.Sorting: is used to establish group order within each partition. In the following example. Change Capture. This is the method the SQL databases use for an ORDER BY clause. the previous input Data Set is partitioned on LName and FName columns. transmitted. a sort is required after repartitioning. we would see the following results: Parallel Framework Red Book: Data Flow Job Design July 17. Instead. . Other stages (for example. FName: LName Dodge Dodge Ford Ford Ford Ford Ford Ford Ford Ford FName John Horace Henry Henry Eleanor Eleanor Edsel Edsel Clara Clara Address 75 Boston Boulevard 17840 Jefferson 66 Edison Avenue 4901 Evergreen 7900 Jefferson 1100 Lakeshore 7900 Jefferson 1100 Lakeshore 66 Edison Avenue 4901 Evergreen However. Merge) require pre-sorted groups of related records. All rights reserved. 2006 74 of 179 © 2006 IBM Information Integration Solutions. Each column is specified with an ascending or descending sort order. For example. Sort Aggregator. To restore row order and groupings. assigning rows with the same key column values to the same partition . one or more secondary key column(s) to generate a sequential.Information Integration Solutions Center of Excellence 6 Sorting Traditionally. The order of key columns determines the sequence and groupings in the result set. or translated into any language in any form by any means without the written permission of IBM. sorting on primary key LName (ascending). as illustrated in the following example. when data is re-partitioned. based on one or more key column(s) NOTE: By definition. Given a 4-node configuration file. the process of sorting data uses one primary key column and. in most cases there is no need to globally sort data to produce a single sequence of rows.Partitioning: is used to gather related records. No part of this publication may be reproduced. ordered result set. the Remove Duplicates stage selects either the first or last row from each group of an input Data Set sorted by one or more key columns. sorting is most often needed to establish order within specified groups of data. sort order is not maintained. secondary key FName (descending): Input Data: I D 1 2 3 4 5 6 7 8 9 10 LName Ford Ford Ford Ford Dodge Dodge Ford Ford Ford Ford FName Henry Clara Edsel Eleanor Horace John Henry Clara Edsel Eleanor Address 66 Edison Avenue 66 Edison Avenue 7900 Jefferson 7900 Jefferson 17840 Jefferson 75 Boston Boulevard 4901 Evergreen 4901 Evergreen 1100 Lakeshore 1100 Lakeshore I D 6 5 1 7 4 10 3 9 2 8 After Sorting by LName.

All rights reserved.4. transmitted. Let’s say that an input Data Set consists of order history based on CustID and Order Date.2 Sorting and Hashing Advanced Example provides a more detailed discussion and example of partitioning and sorting. . sequential ordered result is needed. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Partition 0 ID 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen ID 4 6 10 LName Ford Dodge Ford Partition 2 FName Eleanor John Eleanor Address 7900 Jefferson 75 Boston Boulevard 1100 Lakeshore Partition 1 ID 3 5 9 LName Ford Dodge Ford FName Edsel Horace Edsel Address 7900 Jefferson 17840 Jefferson 1100 Lakeshore Partition 3 ID 1 7 LName Ford Ford FName Henry Henry Address 66 Edison Avenue 4901 Evergreen Applying a parallel sort to this partitioned input Data Set. secondary sort keys can be used to establish order within a group for selection with the Remove Duplicates stage (which can specify First or Last duplicate to retain). To satisfy these requirements we could:  Partition on CustID to group related records  Sort on OrderDate in Descending order  Remove Duplicates on CustID. Using Remove Duplicates. we want to select the most recent order for a given customer. transcribed. using the primary key column LName (ascending) and secondary key column FName (descending) would generate the resulting Data Set: Partition 0 ID 2 8 LName Ford Ford FName Clara Clara Address 66 Edison Avenue 4901 Evergreen ID 6 4 10 LName Dodge Ford Ford Partition 2 FName John Eleanor Eleanor Address 75 Boston Boulevard 7900 Jefferson 1100 Lakeshore Partition 1 ID 5 3 9 LName Dodge Ford Ford FName Horace Edsel Edsel Address 17840 Jefferson 7900 Jefferson 1100 Lakeshore ID 1 7 LName Ford Ford Partition 3 FName Henry Henry Address 66 Edison Avenue 4901 Evergreen Note that the partition and sort keys do not always have to match. 2006 75 of 179 © 2006 IBM Information Integration Solutions.2 Complete (Total) Sort If a single. For example.partition and parallel Sort on key column(s) Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. 6. stored in a retrieval system. in general it is best to use a two step process: . with Duplicate To Retain=First 12.

use the Link sort unless a specific option is needed on the stand-alone Stage. The Link sort offers fewer options. the “Perform Sort” option is checked.Information Integration Solutions Center of Excellence - use a Sort Merge collector on these same key column(s) to generate a sequential. as shown below: Figure 35: Link Sort icon Additional properties can be specified by right-clicking on the key column as shown in the following illustration: Key column options let the developer specify: .case sensitivity (strings) .3 Link Sort and Sort Stage DataStage Enterprise Edition provides two methods for parallel sorts – the standalone sort stage (when execution mode is set to Parallel) and sort on a link (when using a keyed input partitioning method). transcribed. All rights reserved. when specifying a keyed partitioning method. partitioning. transmitted.Sorting character set: ASCII (default) or EBCDIC (strings) Parallel Framework Red Book: Data Flow Job Design July 17. . When specifying key column(s) for partitioning. Within the Designer canvas. ordered result set This is similar to the way parallel database engines perform their parallel sort operations. 6.3. the standalone Sort stage is used to specify the Sort Key mode for partial sorts.key column usage: sorting.sort direction: Ascending or Descending . In general. No part of this publication may be reproduced. (Sorting on a link is not available with Auto partitioning. links that have sort defined will have a Sort icon in addition to the partitioning icon. both methods use the same internal sort package (the tsort operator). 6. stored in a retrieval system. but is easier to maintain in a DataStage job. The stand-alone sort offers more options. or both . By default. as there are fewer stages on the design canvas. although the Enterprise Edition engine may insert a sort if required). 2006 76 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. Most often. but as a separate stage makes job maintenance slightly more complicated.1 Link Sort Sorting on a link is specified on the Input/Partitioning stage options.

the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization. Always specify “DataStage” Sort Utility. the following properties are not available when sorting on a link: Sort Key Mode (a particularly important performance optimization) Create Cluster Key Change Column Create Key Change Column Output Statistics Sort Utility (don’t change this!) Restrict Memory Usage Figure 37: Sort Stage options Of the options only available in the standalone Sort stage. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys. 2006 77 of 179 © 2006 IBM Information Integration Solutions. which is significantly faster than a “UNIX” sort.5 Sub-Sorts Within the standalone Sort stage. transcribed. It is used when resorting a sub-grouping of a previously sorted input Data Set. No part of this publication may be reproduced. Specifically. All rights reserved. the Sort Key Mode is most frequently used. disable Stable sort unless needed.Information Integration Solutions Center of Excellence - Position of nulls in the result set (for nullable columns) Figure 36: Specifying Link Sort options 6. instead of performing a complete Sort.4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. transmitted. - 6.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. It is important to note that by default the Stable sort option is disabled for sorts on a link and Enabled with the standalone Sort stage. This “subsort” uses significantly less disk space and CPU resource.3. NOTE: The Sort Utility option is an artifact of previous releases. This requires some additional overhead in the sort algorithm. stored in a retrieval system. For this reason. 6. and can often be performed in memory (depending on the size of the new subsort groups). or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. .

p1] node3[op2. while new sort keys are specified with the “Sort” key mode property. .Information Integration Solutions Center of Excellence To resort based on a sub-grouping. Remove Duplicates.6 Automatically-Inserted Sorts By default. Enterprise Edition inserts sorts before any stage that requires matched key values or ordered groupings of (Join. all key columns must still be defined in the Sort stage. 6. keys with “Don’t Sort (Previously Sorted)” property must be at the top of the list. And. stored in a retrieval system.p0] node2[op2. transmitted. Sorts are only inserted automatically when the flow developer has not explicitly defined an input sort. or translated into any language in any form by any means without the written permission of IBM. without gaps between them. Re-used sort keys are specified with the “Don’t Sort (Previously Sorted)” property. Merge. 2006 78 of 179 © 2006 IBM Information Integration Solutions. as shown in this score fragment: op1[4p] {(parallel inserted tsort operator {key={value=LastName}. the key column order for these keys must match the key columns and order defined in the previously-sorted input Data Set. as shown in the following example: Figure 38: Sort Key Mode property To successfully perform a subsort. Parallel Framework Red Book: Data Flow Job Design July 17.4. key={value=FirstName}}(0)) on nodes ( node1[op2. No part of this publication may be reproduced.2 Understanding the Parallel Job Score) can be used to identify automatically-inserted sorts. DataStage Enterprise Edition inserts sort operators as necessary to ensure correct results.p3] )} Typically. If the input data does not match the key column definition for a subsort. the job will abort.p2] node4[op2. transcribed. Sort Aggregator). All rights reserved. The parallel job score (see 12.

7 Sort Methodology Using the rules and behavior outlined in the previous section. No part of this publication may be reproduced. There are two ways to prevent Enterprise Edition from inserting an un-necessary sort: a) Insert an upstream Sort stage on each link. transmitted. Create Cluster Key Change Column. On a global basis.Information Integration Solutions Center of Excellence While ensuring correct results. If the sort operation can be performed in memory (as is often the case with a subsort) then no disk I/O is performed. 6. Revisiting the partitioning examples in Section 5. each sort uses 20MB of memory per partition for its memory buffer. inserted sorts can be a significant performance impact if they are not necessary.8 Tuning Sort Sort is a particularly expensive task within DataStage Enterprise Edition which requires CPU. To perform a sort. This will verify sort order but not actually perform a sort. aborting the job if data is not in the required sort order. or translated into any language in any form by any means without the written permission of IBM. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be Parallel Framework Red Book: Data Flow Job Design July 17. Output Statistics  Always specify “DataStage” Sort Utility for standalone Sort stages  Use the “Sort Key Mode=Don’t Sort (Previously Sorted)” to resort a sub-grouping of a previously-sorted input Data Set e) Be aware of automatically-inserted sorts  Set $APT_SORT_INSERTION_CHECK_ONLY to verify but not establish required sort order f) Minimize the use of sorts within a job flow g) To generate a single. the environment variable $APT_SORT_INSERTION_CHECK_ONLY should be set to prevent Enterprise Edition from inserting un-necessary sorts before the Join stage. the following methodology should be applied when sorting in a DataStage Enterprise Edition data flow: a) b) c) d) Start with a link sort Specify only necessary key column(s) Don’t use Stable Sort unless needed Use a stand-alone Sort stage instead of a Link sort for options that not available on a Link sort:  Sort Key Mode. sequential ordered result set use a parallel Sort and a Sort Merge collector 6. Create Key Change Column. rows in the input Data Set are read into a memory buffer on each partition. 2006 79 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. define all sort key columns with the Sort Mode key property “Don’t Sort (Previously Sorted)” b) Set the environment variable APT_SORT_INSERTION_CHECK_ONLY. This value can be changed for each standalone Sort stage using the “Restrict Memory Usage” option (the minimum is 1MB/partition). stored in a retrieval system. By default. transcribed. memory. and disk resources. .4: Partitioning Examples.

.scratch disks defined in the current configuration file (APT_CONFIG_FILE) in the “sort” named disk pool . then results are temporarily spooled to disk in the following order: . transmitted. in MB.the default directory specified by the environment variable TMPDIR . overriding any per-sort specifications. transcribed. stored in a retrieval system.the directory “/tmp” (on UNIX) or “C:/TMP” (on Windows) if available The file system configuration and number of scratch disks defined in parallel configuration file can greatly impact the I/O performance of a parallel sort. Having a greater number of scratch disks for each node allows the sort to spread I/O across multiple file systems. 2006 80 of 179 © 2006 IBM Information Integration Solutions. If the input Data Set cannot fit into the sort memory buffer. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence use to specify the size of the memory buffer. Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM.scratch disks defined in the current configuration file default disk pool . All rights reserved. for all sort operators (link and standalone).

All rights reserved. Can only be written – contents cannot be read or verified.) Rare instances where lookup reference data is required by multiple jobs and is not updated frequently. Can only be read from and written to by DataStage parallel jobs or orchadmin command.Information Integration Solutions Center of Excellence 7 File Stage Usage 7. Slightly higher overhead than Data Set. Limitations Cannot write to a single file in parallel. Data Sets provide maximum performance for reading and writing data from disk. such as mainframe sources with COBOL copybook file definitions. performance penalty of conversion. . transcribed.1 Which File Stage to Use DataStage/EE offers various stages for reading from and writing to files. However. (Requires SAS connectivity license for DataStage. using the internal format of the parallel engine.2 Data Set Usage Parallel Data Sets are the persistent (on-disk) representation of the in-memory data structures of DS/EE. 2006 81 of 179 © 2006 IBM Information Integration Solutions. Intermediate storage between DataStage parallel jobs. performance penalty of format conversion. Data Sets store data in partitioned form. Complex Flat File Data Set File Set SAS Parallel Data Set Lookup File Set No DS/EE file stage supports “update” of existing records. Cannot write in parallel. can write in parallel (generates multiple segment files). Need to share data with an external Parallel SAS application. but this is not recommended as it imposes risks for failure recovery. Can only be used as reference link on a Lookup stage. As such. transmitted. are summarized below: File Stage Recommended Usage Read and write standard files in a single format. In general. Data Sets can only be read from and written to using a DataStage parallel job. stored in a retrieval system. can only be read from / written to by DS/EE or Parallel SAS. Recommendations for when to use a particular stage. No part of this publication may be reproduced. Need to share information with external applications. does not support hierarchical data files. Requires Parallel SAS. as no overhead is needed to translate data to the internal DS/EE representation. If data is Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. Some stages (parallel Data Set) support “Append” to add new records to an existing file. Sequential File Need to read source data in complex (hierarchical) format. and any limitations. 7.

3 Sequential File Stages (Import and Export) The Sequential File stage can be used to read from or write to one or more flat files of the same format. transcribed. 2006 82 of 179 © 2006 IBM Information Integration Solutions. transmitted. The formatting and column properties of the Column Import stage match those of the Sequential File stage. Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. stored in a retrieval system.file may be either fixed or variable-width Read Method: Specific Files. and then pass this to a Column Import stage to parse the file in parallel. Read From Multiple Nodes option is set to Yes useful for cluster and Grid configurations . using the format options specified in the FileSet stage. A better option for writing to a set of Sequential Files in parallel is to use the FileSet stage.2 Writing to a Sequential File in Parallel It is only possible to write in parallel from a Sequential File stage when more than one output file is specified. All rights reserved. in parallel. Readers Per Node option greater than 1 useful for SMP configurations .file may only be fixed-width Note that when reading in parallel. only one file specified.1 Reading from a Sequential File in Parallel The ability to read Sequential File(s) in parallel within Enterprise Edition depends on the Read Method and the options specified: Sequential File – options to read sequentially: Read Method: Specific Files. This will create a single header file (in text format) and corresponding data files. input row order is not maintained across readers. the Sequential File stage can only read and write data that is in flattened (row/column) format.3 Separating I/O from Column Import If the Sequential File input cannot be read in parallel.Information Integration Solutions Center of Excellence 7.3.3. define a single large string column for the non-parallel Sequential File read. The FileSet stage will write in parallel.3. 7. 7. only one file specified may be a file or named pipe Read Method: File Pattern Sequential File – options to read in parallel: Read Method: Specific Files. 7. . set environment variable $APT_IMPORT_PATTERN_USES_FILESET Read Method: Specific Files. performance can still be improved by separating the file I/O from the column parsing operation. the degree of parallelism of the write will correspond to the number of file names specified. more than one file specified each file specified within a single Sequential File stage must be of the same format Read Method: File Pattern. Unlike the Complex Flat File stage. or translated into any language in any form by any means without the written permission of IBM. In these instances. As shown in the following Job fragment.

making the entire downstream flow run sequentially (unless it is later repartitioned). 7. since the Import / Export properties used by the Sequential File and Column Import stages are not documented in the DataStage Parallel Job Developer’s Guide. the buffers are always flushed to disk.7 Reading and Writing Nullable Columns When reading from or writing to Sequential Files or File Sets.3.4 Partitioning Sequential File Reads Care must be taken to choose the appropriate partitioning method from a Sequential File read: • Don’t read from Sequential File using SAME partitioning in the downstream stage! Unless more than one source file is specified. 7. When multiple files are read by a single Sequential File stage (using multiple files.Information Integration Solutions Center of Excellence Figure 39: Column Import example Note that this method is also useful for External Source and FTP sequential source stages. All rights reserved. The environment variable $APT_EXPORT_FLUSH_COUNT allows the job developer to specify how frequently (in number of rows) that the Sequential File stage flushes its internal buffer on writes. Using the Schema File option allows the format of the source file to be specified at runtime. transmitted. When a job completes successfully. SAME will read the entire file into a single partition. including Sequential File import / export format properties is documented in the Orchestrate Record Schema manual. the Sequential File (export operator) stage buffers its writes to optimize performance. . instead of statically through Table Definitions. It is also important to remember that this setting will apply to all Sequential File stages in the data flow. The format of the Schema File. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column.6 Parameterized Sequential File Format The Sequential File stage supports a Schema File option to specify the column definitions and file format of the source file. as shown below: Parallel Framework Red Book: Data Flow Job Design July 17. but there is a small performance penalty associated with increased I/O. or by using a File Pattern).3. It is important to use ROUNDROBIN partitioning (or other partitioning appropriate to downstream components) to evenly distribute the data in the flow. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. Setting this value to a low number (such as 1) is useful for realtime applications. Note that this document is required.5 Sequential File (Export) Buffering By default.3. • 7. stored in a retrieval system. transcribed. each file’s data is read into a separate partition. 2006 83 of 179 © 2006 IBM Information Integration Solutions.3. 7.

To display each field value. Double-click on the column number in the grid dialog to set this column property. Double-click on the column number in the grid dialog or right mouse click on the column and select edit column to set these properties.8 Reading from and Writing to Fixed-Length Files Particular attention must be taken when processing fixed-length fields using the Sequential File stage: • If the incoming columns are variable-length data types (for example. the field width and pad string column properties must be set to match the fixed-width of the output column. Use caution when specifying this option as it can generate an enormous amount of detail in the job log. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. Double-click on the column number in the grid dialog to set this column property. the field width column property must be set to match the fixed-width of the input column. transcribed. • • 7. Integer. Decimal. Decimal. All import and export properties are listed in the Import/Export Properties chapter of the Orchestrate Operators Reference. transmitted. bounded-length Varchar columns (Varchars with the length option set). stored in a retrieval system. 2006 84 of 179 © 2006 IBM Information Integration Solutions.3. use the print_field import property. When writing fixed-length files from variable-length fields (eg. if the source file has fields with values longer than the maximum Parallel Framework Red Book: Data Flow Job Design July 17. By default. . Varchar). • If a field is nullable.3. Integer. you must define the null field value and length in the Nullable section of the column property.Information Integration Solutions Center of Excellence Figure 40: Extended Column Metadata (Nullable properties) 7.9 Reading Bounded-Length VARCHAR Columns Care must be taken when reading delimited. Varchar). No part of this publication may be reproduced.

in some disk array configurations. the stage allows you to write data to one or more complex flat files. or translated into any language in any form by any means without the written permission of IBM. the stage allows you to read data from one or more complex flat files. stored in a retrieval system. 7. but you can configure the stage to execute sequentially if it is only reading one file with a single reader.3. 7. Finally. transmitted. Increasing this size may improve performance. NOTE: The Complex Flat File stage cannot read from sources with OCCURS DEPENDING ON clauses. the environment variables $APT_IMPORT_BUFFER_SIZE and $APT_EXPORT_BUFFER_SIZE can be used to improve I/O performance.4 Complex Flat File Stage The Complex Flat File (CFF) stage can be used to read or write one or more files in the same hierarchical format. (This is an error in the DataStage documentation. transcribed.) When used as a target. It does not write to MVS datasets.10 Tuning Sequential File Performance On heavily-loaded file servers or some RAID/SAN array configurations. When used as a source. Complex Flat File source stages execute in parallel mode when they are used to read multiple files. 2006 85 of 179 © 2006 IBM Information Integration Solutions. or OCCURS clauses. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Varchar length. including MVS datasets with QSAM and VSAM files. A complex flat file may contain one or more GROUPs. The environment variable $APT_IMPORT_REJECT_STRING_FIELD_OVERRUNS will direct Enterprise Edition to reject records with strings longer than their declared maximum column length. with a default of 128 (128K). REDEFINES. These settings specify the size of the read (import) and write (export) buffer size in Kbytes. setting the environment variable $APT_CONSISTENT_BUFFERIO_SIZE to a value equal to the read/write size in bytes can significantly improve performance of Sequential File operations. . these extra characters will be silently truncated. All rights reserved. Parallel Framework Red Book: Data Flow Job Design July 17.

y] or string[x+y] decimal[x+y. native binary binary. leading Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence 7.y] sfloat dfloat ustring[n] ustring[max=n] subrec Internal Options packed packed zoned zoned.y] or string[x+y] decimal[x+y. or translated into any language in any form by any means without the written permission of IBM. 2006 86 of 179 © 2006 IBM Information Integration Solutions.y] decimal[x+y.y] decimal[x+y. leading separate. graphic_g vargraphic_g/n Size 2 bytes 4 bytes 2 bytes 2 bytes 4 bytes 8 bytes n bytes n bytes n bytes (x+y)/2+1 bytes (x+y)/2+1 bytes x+y bytes x+y bytes x+y bytes x+y bytes x+y+1 bytes x+y+1 bytes 4 bytes 8 bytes n*2 bytes n*2 bytes Internal Type int16 int32 int64 uint16 uint32 uint64 string(n) raw(n) string(max=n) decimal[x+y. All rights reserved. native binary binary. stored in a retrieval system. native binary binary. trailing zoned.1 CFF Stage Data Type Mapping When you work with mainframe data using the CFF stage. native binary character character for filler varchar decimal decimal display_numeric display_numeric display_numeric display_numeric display_numeric display_numeric float float graphic_n. native binary binary. .y] decimal[x+y. trailing separate. No part of this publication may be reproduced. the data types are mapped to internal Enterprise Edition data types as follows: COBOL Type S9(1-4) COMP/COMP-5 S9(5-9) COMP/COMP-5 S9(10-18) COMP/COMP-5 9(1-4) COMP/COMP-5 9(5-9) COMP/COMP-5 9(10-18) COMP/COMP-5 X(n) X(n) X(n) 9(x)V9(y)COMP-3 S9(x)V9(y)COMP-3 9(x)V9(y) S9(x)V9(y) S9(x)V9(y) SIGN IS TRAILING S9(x)V9(y) SIGN IS LEADING S9(x)V9(y) SIGN IS TRAILING SEPARATE S9(x)V9(y) SIGN IS LEADING SEPARATE COMP-1 COMP-2 N(n) or G(n) DISPLAY-1 N(n) or G(n) DISPLAY-1 Group Description binary. trailing zoned.y] decimal[x+y.y] decimal[x+y.4. native binary binary. transmitted. transcribed.

by definition. it is important to minimize the number of transformers. transcribed. When rows are rejected by a Transformer. the Transformer will reject the row instead of sending it to the output link(s). For this reason. Because the DS/EE Transformer runs in parallel. stored in a retrieval system.1. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link. For example.8. For this reason. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17.4: Parallel Transformer stages for guidelines on Transformer stage usage. All rights reserved. To create a Transformer reject link in Designer. 2006 87 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence 8 Transformation Languages 8. entries are placed in the Director job log.col) Then “” Else link.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer. and to use other stages (such as Copy) when derivations are not needed. if you intend to use a nullable column within a Transformer derivation or output link constraint. Even if the target column in an output derivation allows nullable results. or translated into any language in any form by any means without the written permission of IBM. See Section 3. it should be converted from its out-of-band (internal) null representation to an in-band (specific value) null representation using stage variables or the Modify stage. No part of this publication may be reproduced. transmitted. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. undefined. 8. This makes it easy to identify reject conditions (by row counts).1. 8. . the Transformer will allow this row to be sent to the output link without being rejected.1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component.col Note that if an incoming column is only used in an output column mapping. Always include reject links in a parallel Transformer.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints. right-click on an output link and choose “Convert to Reject: Figure 41: Transformer Reject link The parallel Transformer rejects NULL derivation results (including output link constraints) because the rules for arithmetic and string handling of NULL values are.

stored in a retrieval system. 1.4 -> -1.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7. and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example.1. transcribed. Within the link constraints dialog box. transmitted.5-> 2. there are times when internal decimal variables need to be generated in order to perform the evaluation. Since the Transformer will abort the entire job flow immediately. these internal decimal variables will have a precision and scale of 38 and 10. For example. Therefore. internal decimal results are rounded to the nearest applicable value.4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule. By default. 1). or translated into any language in any form by any means without the written permission of IBM. the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. apply the abort rule to this output link. -1. When generating a sequence of numbers in parallel.1. No part of this publication may be reproduced. Thus. the PadString function uses the length of the source type.6 ->1. 8. -1. up to a maximum precision of 255 and scale of 125. -1. For example. If more precision is required. Create a new output link that will handle rows that match the abort rule. it is possible that valid rows will not have been flushed from Sequential File (export) buffers. Examples: 1. or performing parallel derivations.1. Examples: 1. The environment variable APT_DECIMAL_INTERM_ROUND_MODE can be used to change the rounding behavior using one of the following keywords: ceil Rounds towards positive infinity.3.5: Sequential File (Export) Buffering). or committed to database tables. -1. it is important to make sure the type conversion is done before a row reaches the Transformer. By default.4 -> -2 round_inf Rounds or truncates towards nearest representable value.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition.6 -> -1 floor Rounds towards negative infinity.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment. 8.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17. breaking ties by rounding positive values toward positive infinity and negative values toward negative infinity. not the target. TrimLeadingTrailing(string) works only if string is a VarChar field. 2006 88 of 179 © 2006 IBM Information Integration Solutions. the incoming column must be type VarChar before it is evaluated in the Transformer.4 -> 2.4 -> 1. 8. Examples: 1. All rights reserved. . the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range.

it is useful to understand what items get evaluated and when. This can be made more efficient by moving the substring calculation into a stage variable.. The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value.. unless the derivation is empty For each output link: Evaluate the link constraint.56 ->-1.5. From this sequence. the substring is evaluated just once for every input row. Similarly. then the following test may appear in a number of output columns derivations: IF (DSLINK1. By doing this. suppose multiple columns in output links want to use the same substring of an input column. . Examples: 1. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example.col[1. For example. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. 8. if true Evaluate each column derivation value Write the output record Else skip the link Next output link Next input row The stage variables and the columns within a link are evaluated in the order in which they are displayed in the Transformer editor.1. the output links are also evaluated in the order in which they are displayed. -1.56 -> 1. 2006 89 of 179 © 2006 IBM Information Integration Solutions.3] = “001”) THEN . All rights reserved. the stage variable definition would be: Parallel Framework Red Book: Data Flow Job Design July 17.5. round or truncate to the scale size. transmitted. transcribed. In this case. the evaluation of the substring of DSLINK1.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. it can be seen that there are certain constructs that would be inefficient to include in output column derivations.col[1. as they would be evaluated once for every output column that uses them.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign.3] is evaluated for each column that uses it. stored in a retrieval system. In this case.

a column definition may include a function call that returns a constant value. The stage variable would be: IF (DSLink1. the function would still be evaluated once for every input row.3] and each column derivation would start with: IF (StageVar1 = “001” THEN . its value for the whole Transformer processing is unchanged from the initial value. Any expression that previously used this function would be changed to use the stage variable instead.Information Integration Solutions Center of Excellence DSLINK1. In this case. another example would be part of an expression such as: Parallel Framework Red Book: Data Flow Job Design July 17. Therefore. or translated into any language in any form by any means without the written permission of IBM.20) This returns a string of 20 spaces. but in this case. This can be achieved using stage variables. This function could be moved into a stage variable derivation.20) You would then leave the derivation of the stage variable on the main Transformer page empty... this example could be improved further by also moving the string comparison into the stage variable. Then. In addition to a function value returning a constant value. A stage variable can be assigned an initial value from the Stage Properties dialog/Variables tab in the Transformer stage editor. The initial value of the stage variable is evaluated just once.col1[1. In this case. The solution here is to move the function evaluation into the initial value of a stage variable. transcribed. 2006 90 of 179 © 2006 IBM Information Integration Solutions. In fact. such as: Str(“ “. before any input rows are processed. All rights reserved. . because the derivation expression of the stage variable is empty. the variable would have its initial value set to: Str(“ “. transmitted.col[1. It would be more efficient to calculate the constant value just once for the whole Transformer. • Where an expression includes calculated constant values For example. it is not reevaluated for each input row.3] = “001” THEN 1 ELSE 0 and each column derivation would start with: IF (StageVar1) THEN This reduces both the number of substring functions evaluated and string comparisons made in the Transformer. the function would be evaluated every time the column derivation is evaluated. stored in a retrieval system. No part of this publication may be reproduced.

Information Integration Solutions Center of Excellence "abc" : "def" As with the function-call example. transmitted. . again. All rights reserved. If this just appeared once in one output column expression. The Modify stage uses the syntax of the underlying modify operator. Parallel Framework Red Book: Data Flow Job Design July 17. Since the subpart of the expression is actually constant.col1+"1" In this case. this constant part of the expression could again be moved into a stage variable. stored in a retrieval system. where that conversion would have been required. or translated into any language in any form by any means without the written permission of IBM. Otherwise. and then use the stage variable in place of DSLink1. if DSLINK1. then it would be more efficient to use a stage variable to perform the conversion once. since it uses low-level functionality that is part of every DataStage Enterprise Edition component. and so.col1+1 In this example. specify its derivation to be DSLINK1. then. In this case.2 Modify Stage The Modify stage is the most efficient “stage” available.col1. However. the Output Mapping properties for any parallel stage will generate an underlying modify for default data type conversions. needless conversions are required wherever that variable is used. an expression may include something like this: DSLink1. an integer stage variable. using the initial value setting to perform the concatenation just once. As noted in the previous section. in order to be able to add it to DSLink1. this concatenation is evaluated every time the column derivation is evaluated. transcribed. For example. a conversion would be required every time the expression is evaluated. null conversion. where it requires the same type conversion in each expression. It should be noted that when using stage variables to evaluate parts of expressions. the data type of the stage variable should be set correctly for that context. The standalone Modify stage can be used for non-default type conversions (nearly all date and time conversions are non-default). the "1" is a string constant. documented in the Parallel Job Developers Guide as well as the Orchestrate Operators Reference. dropping and renaming columns. for example. it must be converted from a string to an integer each time the expression is evaluated.col1 were a string field. you would create. The solution in this case is just to change the constant from a string to an integer: DSLink1. and string trim. or it is used in multiple places. this would be fine.col1. • Where an expression requiring a type conversion is used as a constant. 2006 91 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. if an input column is used in more than one expression. 8.col1.

2. By default. The character argument is the character to remove.dataType is its optional data type. No part of this publication may be reproduced. . By default.value is the value you wish to represent a null in the output. stored in a retrieval system. All rights reserved. these characters are retained when the fixed-length string is then converted back to a variable-length string. use it if you are also converting types. string. 8. or translated into any language in any form by any means without the written permission of IBM. for decimal. . transcribed.sourceField is the source field’s name.value) where: . 2006 92 of 179 © 2006 IBM Information Integration Solutions. To convert from an in-band null to an out-of-band null. For a numeric field value can be a numeric value. NOTE: The DataStage Parallel Job Developers Guide gives incorrect syntax for converting an out-of-band null to an in-band null (value) representation.dataType is its optional data type. The following example removes all leading ASCII NULL characters from the beginning of name and places the remaining characters in an output variable-length string with the same name: Parallel Framework Red Book: Data Flow Job Design July 17.value) where: . with the following syntax: stringField=string_trim[character.destField is the destination field’s name. this is NULL. transmitted. the syntax is: destField[:dataType] = handle_null (sourceField. To convert from an out-of-band null to an in-band null (value) representation within Modify. the syntax is: destField[:dataType] = make_null(sourceField. and justify defaults to begin. justify] (string) You can use this function to remove the characters used to pad variable-length strings when they are converted to fixed-length strings of greater length.2. . date. direction.2 Modify and String Trim The function string_trim has been added to Modify. . time. value can be a string. The destField is converted from an Orchestrate out-of-band null to a value of the field’s data type. and timestamp fields.1 Modify and Null Handling The Modify stage can be used to convert an out-of-band null value to an in-band null representation and vice-versa. Justify has no affect when the target string has variable length.sourceField is the source field’s name .destField is the destination field’s name.Information Integration Solutions Center of Excellence 8. use it if you are also converting types. direction defaults to end.value is the value of the source field when it is null. . The value of the direction and justify arguments can be either begin or end. .

transmitted. No part of this publication may be reproduced. All rights reserved. . 2006 93 of 179 © 2006 IBM Information Integration Solutions. end. begin](name) The following example removes all trailing Z characters from color. transcribed.Information Integration Solutions Center of Excellence name:string = string_trim[NULL. or translated into any language in any form by any means without the written permission of IBM. and left-justifies the resulting hue fixed-length string: hue:string[10] = string_trim[‘Z’. stored in a retrieval system. begin](color) Parallel Framework Red Book: Data Flow Job Design July 17.

This is also handy with Lookups that have multiple reference links. No part of this publication may be reproduced. and ODBC Enterprise stages) to scenarios where the number of input rows is significantly smaller (for example. Oracle Enterprise.1 Lookup vs. Join).2 Capturing Unmatched Records from a Join The Join stage does not provide reject handling for unmatched records (such as in an InnerJoin scenario). In an OUTER join scenario. It is best to test both the Sparse and Normal to see which actually performs best. the JOIN or MERGE stage should be used. stored in a retrieval system.Information Integration Solutions Center of Excellence 9 Combining Data 9. Limit the use of database Sparse Lookups (available in the DB2 Enterprise. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 94 of 179 © 2006 IBM Information Integration Solutions. . If un-matched rows must be captured or logged. Join vs. or translated into any language in any form by any means without the written permission of IBM. This isolates your match/no-match logic from any changes in the metadata. If the Data Sets are larger than available memory resources. A Transformer stage can be used to test for NULL values in unmatched columns. the default value for a Varchar is an empty string (“”). In most cases. During an Outer Join. 9. 1:100 or more) than the number of reference rows.7: Database Sparse Lookup vs. the default value for an Integer is zero. (see Section 10. Left Outer. then Enterprise Edition inserts “default” values based on the data type. All rights reserved. with a constant value. to each of the inner links and test that column for the constant after you have performed the join. This is most easily done by inserting a Copy stage and mapping a column from NON-NULLABLE to NULLABLE. For this reason. transmitted. and to retest if the relative volumes of data change dramatically. when a match does not occur. transcribed. and the default value for a Char is a string of padchar characters equal to the length of the Char column. or both links in the case of Full Outer) are output regardless of match on key values. Each lookup reference requires a contiguous block of shared memory. all rows on an outer link (for example. For example.1. an OUTER join operation must be performed. Sparse Lookups may also be appropriate for exception-based processing when the number of exceptions is a small fraction of the main input data. it is best to use a Column Generator to add an ‘indicator’ column. care must be taken to change the column properties to allow NULL values before the Join. Merge The Lookup stage is most appropriate when the reference data for all lookup stages in a job is small enough to fit into available physical memory. the Join stage inserts values into the unmatched non-key column(s) using the following rules: a) If the non-key column is defined as nullable (on the Join input links) then Enterprise Edition will insert NULL values in the unmatched columns b) If the non-key column is defined as not-nullable. Right Outer.

3. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. the output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double). Unlike the Hash Aggregator. 9. 2006 95 of 179 © 2006 IBM Information Integration Solutions. use the stages shown on the right to: generate a single constant-value key column using the Column Generator or an upstream Transformer aggregate in parallel on the generated column (partition Round Robin. which maintains the results of each keycolumn value/aggregation pair in memory. No part of this publication may be reproduced. the Sort Aggregator requires presorted data. . 9. To aggregate in decimal precision. stored in a retrieval system.1 Aggregation Method By default.2 Aggregation Data Type By default. transmitted.3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings. but only maintains the calculations for the current group in memory. aggregate on generated key column) there is no need to sort or hash-partition the input data with only one key column value aggregate Sequentially on the generated column - - Note that in this example use two Aggregators are used to prevent the sequential aggregation from disrupting upstream processing.Information Integration Solutions Center of Excellence 9. set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. or translated into any language in any form by any means without the written permission of IBM. The Sort Aggregation Method should be used when the number of key values is unknown or very large. the default Aggregation Method is set to Hash. Because each key value/aggregation requires approximately 2K of memory. the Hash Aggregator should only be used when the number of distinct key values is small and finite. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property.3 The Aggregator Stage 9. Note that performance is typically better if you let calculations occur in floating point (Double) data type and convert the results to decimal downstream in the flow.3. transcribed. An exception to this is financial calculations which should be done in decimal to preserve appropriate precision. To perform a total aggregation.3.

You may need to customize the palette to add hidden stages.1 Database stage types DataStage Enterprise Edition offers database connectivity through native parallel and plug-in stage types. For some databases (DB2. transmitted.1. and Teradata). 2006 96 of 179 © 2006 IBM Information Integration Solutions.1 Database development overview This section is intended to provide guidelines appropriate to accessing any database within DataStage Enterprise Edition. stored in a retrieval system. Subsequent sections provide database-specific tips and guidelines.Information Integration Solutions Center of Excellence 10 Database Stage Guidelines 10. Plug-In Database Stages Dynamic RDBMS DB2/UDB API DB2/UDB Load Informix CLI Informix Load Informix XPS Load Oracle OCI Load RedBrick Load Sybase IQ12 Load Sybase OC Teradata API Teradata MultiLoad (MultiLoad) Parallel Framework Red Book: Data Flow Job Design July 17. Informix. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. . transcribed. Oracle. multiple stage types are available: Teradata MultiLoad (TPump) Native Parallel Database Stages DB2/UDB Enterprise Informix Enterprise ODBC Enterprise Oracle Enterprise SQL Server Enterprise Teradata Enterprise NOTE: Not all database stages (for example. 10. Teradata API) are visible in the default DataStage Designer palette.

1. the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server. . the ODBC Enterprise stage cannot interface with database-specific parallel load technologies. and cannot span multiple servers in a clustered or Grid configuration. plug-in database stages match columns by order. attributes). native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform. the native parallel stages often have more stringent connectivity requirements than plug-in stages. scalability. The benefit of ODBC Enterprise stage comes from the large number of included and third party ODBC drivers to enable connectivity to all major database platforms. Because of their tight integration with database technologies. This allows Enterprise Edition to match return columns by name. operating system. Because there are exceptions to this rule (especially with Teradata). they should only be used when it is not possible to use a native parallel stage. transmitted.1.1.1. 2006 97 of 179 © 2006 IBM Information Integration Solutions. so Table Definitions must match the order of columns in a query. care must be taken to assign the correct data types in the job design. From a design perspective. All rights reserved. For example. the ODBC Enterprise stage cannot read in parallel (although a patch to allow parallel read may be available on some platforms through IBM IIS Support). the DataDirect ODBC Driver Manager is also included. Because plug-in stage types cannot read in parallel. stored in a retrieval system. No part of this publication may be reproduced. 10. types.2 ODBC Enterprise stage In general. However. On UNIX. and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied. 10.Information Integration Solutions Center of Excellence 10. not position in the stage Table Definitions. Native parallel stages always pre-query the database for actual runtime metadata (column names.1 Native Parallel database stages In general. or translated into any language in any form by any means without the written permission of IBM. Parallel Framework Red Book: Data Flow Job Design July 17. not name. Furthermore. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect.1. for maximum parallel performance. Unlike the database-specific parallel stages. and version.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. specific guidelines of when to use various stage types are provided in the database-specific topics in this section. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform.1. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. transcribed.

1. 10. use orchdbutil to import metadata to avoid type conversion issues. transcribed. regardless of the table definitions assigned by the DataStage developer. the DS/EE native parallel database stages always “pre-query” the database source or target to determine the actual metadata (column names. care must be taken that the column data types defined by the DataStage developer match the data types generated by the database stage at runtime.2.columns of the database row correspond to columns of a DS/EE record . However. All rights reserved.1. stored in a retrieval system. Database-specific data type mapping tables are included in the following sections. Informix Enterprise or Oracle Enterprise stages.both DS/EE and relational databases support null values. and selecting “Import from Database Table” option in the wizard as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column The actual metadata used by a DS/EE native parallel database stage is always determined at runtime. data types.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types . or translated into any language in any form by any means without the written permission of IBM.2.1 Runtime metadata At runtime.2 Metadata Import When using the native parallel DB2 Enterprise.rows of the database result set correspond to records of a DS/EE Data Set . This utility is available as a server command line utility and within Designer and Manager using “Import Orchestrate Schema Definitions”. transmitted. 2006 98 of 179 © 2006 IBM Information Integration Solutions. This allows the database stages to match return values by column name instead of position.1. nullability) and partitioning scheme (in some cases) of the source or target table.2 Database Metadata 10. No part of this publication may be reproduced. . For each native parallel database stage: .Information Integration Solutions Center of Excellence 10.

For “Table” read method. For “Auto-Generated” SQL. there may be cases where user-defined functions or logic need to be executed on the database server. always specify the “Select List” subproperty. All rights reserved. the following SQL assigns the alias Total to the calculated column: SELECT store_name. it is best to explicitly specify column names on all source database stages. transmitted. instead of using an unqualified “Table” or SQL “SELECT *” read. If the connection is successful. For example. similar to the illustration on the right: Parallel Framework Red Book: Data Flow Job Design July 17.1.3 Optimizing Select Lists For best performance and optimal memory usage. stored in a retrieval system. SUM(sales) Total FROM store_info GROUP BY store_name Note that in many cases it may be more appropriate to aggregate using the Enterprise Edition Aggregator stage. 10.Information Integration Solutions Center of Excellence Figure 42: orchdbutil metadata import One disadvantage to the graphical orchdbutil metadata import is that the user interface requires each table to be imported individually. The alias name(s) should then be added to the Table Definition within DataStage.1. the DataStage Designer will automatically populate the select list based on the stage’s output column definition. orchdbutil can be scripted to automate the process of importing a large number of tables. The only exception to this rule is when building dynamic database jobs that use runtime column propagation to process all columns in a source table. No part of this publication may be reproduced. transcribed.4 Testing Database Connectivity The “View Data” button on the Output / Properties tab of source database stages lets you verify database connectivity and settings without having to create and run a job. 10. When importing a large number of tables. However.2. As a command. you will see a window with the result columns and data.3 Defining Metadata for Database Functions When using database functions within a SQL SELECT list in a Read or Lookup. or translated into any language in any form by any means without the written permission of IBM.1. it is important to use SQL aliases to explicitly name the calculated columns so that they can be referenced within the DataStage job. Test the connection using View Data button. . 2006 99 of 179 © 2006 IBM Information Integration Solutions. it will be easier to use the corresponding orchdbutil command-line utility from the DataStage server machine. 10.

and you will be prompted to view additional detail. 2006 100 of 179 © 2006 IBM Information Integration Solutions. Clicking YES will display a detailed dialog box with the specific error messages generated by the database stage that can be very useful in debugging a database connection failure.6 Database OPEN and CLOSE Commands The native parallel database stages provide options for specifying OPEN and CLOSE commands. constraints. it is important to separate the transformation process from the database write (Load or Upsert) operation. Figure 44: View Additional Error Detail 10. All rights reserved. depending on data volume) can be used to identify existing rows before they are inserted into the target table. it is not a good idea to let DataStage generate target tables unless they are used for temporary storage.1. logging. For example. . the OPEN command can be used to create a target table. including databasespecific options (tablespace.1. transcribed. After transformation. an error message may appear.5 Designing for Restart To enable restart of high-volume jobs. There are limited capabilities to specify Create table options in the stage. 10. etc) not possible with the “Create” option. In general. and the CLOSE command could be used to select all rows from the temporary table and insert into a final target table. the results should be landed to a parallel Data Set. Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence Figure 43: Sample View Data Output If the connection fails. No part of this publication may be reproduced. These options allow commands (including SQL) to be sent to the database before (OPEN) or after (CLOSE) all rows are read/written/loaded to the database. the OPEN command could be used to create a temporary table. or translated into any language in any form by any means without the written permission of IBM. transmitted. Subsequent job(s) should read this Data Set and populate the target table using the appropriate database stage and write method. OPEN and CLOSE are not offered by plugin database stages. As another example. and doing so may violate data-management (DBA) policies. As a further optimization. a Lookup stage (or Join stage. stored in a retrieval system.

No part of this publication may be reproduced. there may be “gray areas” where the decision should be made based on factors such as developer productivity. By default. transmitted. Join Data read by any database stage can serve as the reference input to a Lookup operation.Information Integration Solutions Center of Excellence It is important to understand the implications of specifying a user-defined OPEN and CLOSE command.8 Appropriate Use of SQL and DataStage When using relational database sources. it is faster to use a DataStage JOIN stage between the input and DB2 reference data than it is to perform a “Sparse” Lookup. and ongoing application maintenance costs. The following guidelines can assist with the appropriate use of SQL and DataStage technologies in a given job flow: • When possible. metadata capture and re-use. the optimal implementation involves leveraging the strengths of each technology to provide maximum throughput and developer productivity. All rights reserved.7 Database Sparse Lookup vs. Further details are outlined in the respective database sections of the Orchestrate Operators Reference which is part of the Orchestrate OEM documentation. a default OPEN statement places a shared lock on the source. this reference data is loaded into memory like any other reference link (“Normal” Lookup). This minimizes impact on network and memory resources. this lock is not sent – and should be specified explicitly if appropriate. 2006 101 of 179 © 2006 IBM Information Integration Solutions. . there is often a functional overlap between SQL and DataStage functionality. In most cases. and leverages the database capabilities. when reading from DB2. 10. 1:100 or more) than the number of reference rows in a DB2 or Oracle table.1. a Sparse Lookup may be appropriate. When directly connected as the reference link to a Lookup stage. IMPORTANT: The individual SQL statements required by a “Sparse” Lookup are an expensive operation from a performance perspective. stored in a retrieval system. sending individual SQL statements to the reference database for each incoming Lookup row. or translated into any language in any form by any means without the written permission of IBM. transcribed. with no intermediate stages. ODBC Enterprise. Sparse Lookup is only available when the database stage is directly connected to the reference link.1. and Oracle Enterprise stages allow the lookup type to be changed to “Sparse”. use a SQL filter (WHERE clause) to limit the number of rows sent to the DataStage job. Parallel Framework Red Book: Data Flow Job Design July 17. 10. Although it is possible to use either SQL or DataStage to solve a given business problem. For scenarios where the number of input rows is significantly smaller (for example. When specifying a user-defined OPEN command. While there are extreme scenarios when the appropriate technology choice is clearly understood. the DB2/UDB Enterprise. For example.

. For maximum scalability and parallel performance.Information Integration Solutions Center of Excellence • Use a SQL Join to combine data from tables with a small number of rows in the same database instance. • • Parallel Framework Red Book: Data Flow Job Design July 17. A join that reduces the result set significantly is also often appropriate to do in the database. When combining data from very large tables. the efficiency of the Enterprise Edition Sort and Join stages can be significantly faster than an equivalent SQL query. or when the source includes a large number of database tables. Avoid the use of database stored procedures (for example. it is best to implement business rules using native parallel DataStage components. In this scenario. stored in a retrieval system. Oracle PL/SQL) on a per-row basis within a high-volume data flow. 2006 102 of 179 © 2006 IBM Information Integration Solutions. transmitted. it can still be beneficial to use database filters (WHERE clause) if appropriate. especially when the join columns are indexed. All rights reserved. transcribed. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM.

1 DB2 Stage Types DataStage Enterprise Edition provides access to DB2 databases using one of 5 stages.2. summarized in the following table: DataStage Stage Name DB2/UDB Enterprise DB2/UDB API DB2/UDB Load ODBC Enterprise Dynamic RDBMS Stage Type Native Parallel Plug-In Plug-In Native Plug-In DB2 Requirement DPF. the “client authentication” configuration of a remote instance is not recommended. 2 It is possible to connect the DB2 UDB stage to a remote database by simply cataloging the remote database in the local instance and then using it as if it were a local database.1 DB2/UDB Enterprise stage Enterprise Edition provides native parallel read. No part of this publication may be reproduced. transcribed. or translated into any language in any form by any means without the written permission of IBM. Check with IBM IIS Support for availability. The DB2/UDB Enterprise stage requires DB2 Enterprise Server Edition on UNIX with Data Partitioning Facility (DPF) option. This will only work when the authentication mode of the database on the remote instance is set to “client authentication”. Parallel Framework Red Book: Data Flow Job Design July 17. 3 A patched version of the ODBC Enterprise stage allowing parallel read is available from IBM IIS Support for some platforms. stored in a retrieval system. you may experience data duplication when working in partitioned instances since the node configuration of the local instance may not be the same as the remote instance. (Before DB2 v8.1. For this reason. and load capabilities to parallel DB2 databases on UNIX using the native parallel DB2/UDB Enterprise stage. .2 DB2 Guidelines 10. this was also called “DB2 EEE”.) Furthermore. All rights reserved. lookup. consult the DataStage documentation (DataStage Parallel Job Developers Guide. If you use the stage in this way. transmitted.2.Information Integration Solutions Center of Excellence 10. upsert. 2006 103 of 179 © 2006 IBM Information Integration Solutions. the DB2 hardware/UNIX/software platform must match the hardware/software platform of the DataStage ETL server. same platform as ETL server 2 Any DB2 via DB2 Client or DB2-Connect Subject to DB2 Loader Limitations Any DB2 via DB2 Client or DBE-Connect Any DB2 via DB2 Client or DB2-Connect Supports Partitioned DB2? Yes / directly to each DB2 node Yes / through DB2 node 0 No Yes / through DB2 node 0 Yes / through DB2 node 0 Parallel Read? Yes Parallel Write? Yes Parallel Sparse Lookup Yes SQL Open / Close Yes No No No3 No Possible Limitations No No Possible Limitations No No No No No No No No For specific details on the stage capabilities. and DataStage Plug-In guides) 10.

or translated into any language in any form by any means without the written permission of IBM. the ability to write in parallel may be limited by the table and index configuration set by the D2 database administrator. parallel component the DB2/UDB Enterprise stage is designed for maximum performance and scalability. Furthermore. The DB2/API (plug-in) stage should only be used to read from and write to DB2 databases on nonUNIX platforms (such as mainframe editions through DB2-Connect). No part of this publication may be reproduced.2. when used as data sources. .2 ODBC and DB2 Plug-In Stages The ODBC Enterprise and plug-in stages are designed for lower-volume access to DB2 databases without the DPF option installed (prior to v8. 10. it may be possible to write to a DB2 target in parallel. databases on UNIX platforms that differ from the platform of the DataStage ETL server. including direct communication with each DB2 database node. 10. By facilitating flexible connectivity to multiple types of remote DB2 database servers. or DB2 databases on Windows or Mainframe platforms (except for the “Load” stage against a mainframe DB2 instance which is not supported). 2006 104 of 179 © 2006 IBM Information Integration Solutions. and reading from . Sparse Lookup is not supported through the DB2/API stage. using the same data partitioning as the referenced DB2 tables. “DB2 EE”).2. transmitted. Because each plug-in invocation will open a separate connection to the same target DB2 database table. plug-in stages cannot read from DB2 in parallel. All rights reserved.2 Connecting to DB2 with the DB2/UDB Enterprise Stage Create a Parallel job and add a DB2/UDB Enterprise stage. These goals are achieved through tight integration with the DB2 RDBMS.writing to DB2 in parallel (where appropriate). Using the DB2/UDB API stage or the Dynamic RDBMS stage. These stages also provide connectivity to non-UNIX DB2 databases. since the DS/EE framework will instantiate multiple copies of these stages to handle the data that has already been partitioned in the parallel framework. the use of DataStage plug-in stages will limit overall performance and scalability. stored in a retrieval system.1. Add the following properties: Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence As a native. transcribed.

this is required for a remote connection in order to retrieve the catalog information from the local instance of DB2 and thus must have privileges for that local instance. transcribed. DataStage assumes you require remote connection. 2006 105 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Single stream . it will try to use it for each of the connections in the job.3 Configuring DB2 Multiple Instances in One DataStage Job Although it is not officially supported. Since a db2nodes. you will need to set the following properties on the DB2/UDB Enterprise stage in your parallel job: • Client Instance Name. Enter the password for connecting to DB2.cfg file can July 17. All rights reserved. • Client Alias DB Name. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. • Server. • User. this is required for a remote connection in order to retrieve the catalog information from the local instance of DB2 and thus must have privileges for that local instance. transmitted. • Database. Otherwise use the DB2 environment variable. Set this to the DB2 client instance name.Two Instances Only reading from one instance and writing to another instance with no other DB2 instances (not sure how many stages of these 2 instances can be added to the canvas for this configuration for lookups) 2. If you set this property. Multiple Stream with N DB2 sources with no DB2 targets reading from 1 to n DB2 instances in separate source stages with no downstream other DB2 stages In order to get this configuration to work correctly. 10. to identify the instance name of the DB2 server. Otherwise use the environment variables $APT_DBNAME or $APT_DB2DBDFT to identify the database. DB2INSTANCE. • Password. you must adhere to all of the directions specified for connecting to a remote instance AND the following: • You must not set the APT_DB2INSTANCE_HOME environment variable.Information Integration Solutions Center of Excellence Figure 45: DB2/UDB Enterprise stage properties For connection to a remote DB2/UDB instance. Your job must meet one of the following configurations (note: the use of the word “stream” refers to a contiguous flow of one stage to another within a single job): 1. Set this to the DB2 client’s alias database name for the remote DB2 server database. . No part of this publication may be reproduced. This is required only if the client’s alias is different from the actual name of the remote server database.2. Optionally set this to the instance name of the DB2 server. Optionally set this to the remote server database name. it is possible to connect to more than one DB2 instance within a single job. Two Stream – One Instance per Steam reading from instance A and writing to instance A and reading from instance B and writing to instance B (not sure how many stages of these 2 instances can be added to the canvas for this configuration for lookups) 3. Enter the user name for connecting to DB2. Once this variable is set.

if the third DB2 column is named 7dig. Dynamic RDBMS) may be used to access data in other instances. To connect to multiple DB2 instances. if the data volumes are sufficiently small. the Enterprise Edition column will be named “APT_37dig” . All rights reserved. 2006 106 of 179 © 2006 IBM Information Integration Solutions. create a sqllib subdirectory and place the remote instance’s db2nodes. Depending on platform configuration and I/O subsystem performance. This converts these special characters into an internal representation that DataStage can understand.the name must start with a letter or underscore character . 10. DS will default to this directory to find the configuration file for the remote instance.the name is case insensitive When there is an incompatibility. the character is replaced by two underscore characters 10. Or.if the DB2 column name does not begin with a letter or underscore.2. DB2 Load. but have the following restrictions: . For example. Enterprise Edition converts the DB2 column name as follows: . DB2 plug-in stages (DB2 API. the string “APT__column#” (two underscores) is added to beginning of the column name.the name can contain only alphanumeric and underscore characters . stored in a retrieval system. transcribed. this will create problems.if the DB2 column name contains a character that is not alphanumeric or an underscore. .4 DB2/UDB Enterprise stage Column Names At runtime. No part of this publication may be reproduced. DB2 Load. Observe the following guidelines when $DS_ENABLE_RESERVED_CHAR_CONVERT is set: Parallel Framework Red Book: Data Flow Job Design July 17. you must build a user on the DataStage server with the same name as the instance you are trying to connect to (the default logic for the DB2/UDB Enterprise stage is to use the instance’s home directory as defined for the UNIX user with the same name as the DB2 instance). we recommend using separate jobs with their respective DB2 environment variable settings. In the users UNIX home directory. although this incurs the overhead of Sequential File stage (corresponding export/import operators) which does not run in parallel. landing intermediate results to a parallel Data Set. set the environment variable $DS_ENABLE_RESERVED_CHAR_CONVERT if your DB2 database uses the reserved characters # or $ in column names.cfg. the native parallel DB2/UDB Enterprise stage translates column names exactly except when a component of a DB2 column name is not compatible with Enterprise Edition column naming conventions. separate jobs can communicate through named pipes. and Dynamic RDBMS plug-in stages. transmitted.2. where column# is the number of the column. or translated into any language in any form by any means without the written permission of IBM.cfg there. which place no limit on the length of a column name.5 DB2/API stage Column Names When using the DB2/API.Information Integration Solutions Center of Excellence only contain information for one instance. Since the APT_DB2INSTANCE_HOME is not set. • In order for DataStage to locate the db2nodes.

stored in a retrieval system. as shown in the following table. the internal column names are displayed rather than the original DB2 names both in table definitions and in the Data Browser. examine the DDL for each schema to be accessed.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour.r) NVARCHAR(n.s] where p is the precision and s is the scale dfloat dfloat int32 decimal string[n] or ustring[n] string[max=n] or ustring[max=n] sfloat int32 sfloat int16 string[max=n] or ustring[max=n] IMPORTANT: DB2 data types that are not listed in the above table cannot be used in the DB2/UDB Enterprise stage. All rights reserved. the result is a time field.6 DB2/UDB Enterprise stage Data Type Mapping The DB2 database schema to be accessed must NOT have any columns with User Defined Types (UDTs). Once the table definition is loaded.r) DATE DATETIME Enterprise Edition Data Type string[n] or ustring[n] string[max=n] or ustring[max=n] date Time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component. .s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n. Use the “db2 describe table [table-name]” command on the DB2 client for each table to be accessed to determine if UDTs are in use. or translated into any language in any form by any means without the written permission of IBM. 10. DECIMAL[p. They are also used in derivations and expressions. Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. The DB2/UDB Enterprise stage converts DB2 data types to Enterprise Edition data types. 2006 107 of 179 © 2006 IBM Information Integration Solutions. however. transcribed. No part of this publication may be reproduced. transmitted.Information Integration Solutions Center of Excellence - Avoid using the strings __035__ and __036__ in your DB2 column names (these are used as the internal representations of # and $ respectively) Import meta data using the Plug-in Meta Data Import tool. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. DB2 Data Type CHAR(n) CHARACTER VARYING(n. and you should use them if entering SQL in the job yourself. decimal[p. Alternatively.2. avoid hand editing (this minimizes the risk of mistakes or confusion). The original names are used in generated SQL statements. the result is a timestamp field.

performing lookups against. b) The DB2 Load method requires that the DataStage user running the job have DBADM privilege on the target DB2 database. it may be beneficial to have the DB2 DBA configure separate DB2 coordinator nodes (no local data) on each ETL server (in clustered ETL configurations).run in Truncate mode to clear the load pending state. transcribed. reading from and writing to DB2 in parallel (where appropriate). When writing to a DB2 database in parallel. the DB2/UDB Enterprise stage offers the choice of SQL (insert / update / upsert / delete) or fast DB2 loader methods. when using user-defined SQL without partitioning against large volumes of DB2 data. • DB2/UDB Enterprise stage is tightly integrated with the DB2 RDBMS. The DB2 load operator performs a non-recoverable load. That is. parallel component. In these instances. All rights reserved. In this scenario.8 Performance Notes In some cases. the overhead of routing information through a remote DB2 coordinator may be significant.2.9 DB2 in the DataStage USS environment The manner in which DataStage / USS Edition interfaces with DB2 is slightly different than it is in the non-z/OS environment. and using the same data partitioning as the referenced DB2 tables. the DB2 Load DataStage job must be re.2. All operations are logged to the DB2 database log. if the load operation is terminated before it is completed. 2006 108 of 179 © 2006 IBM Information Integration Solutions. In this configuration. and recoverability. During the load operation. Parallel Framework Red Book: Data Flow Job Design July 17. database log usage. 10. 10. All activity in the z/OS environment always goes through the DB2 coordinator node so parallelism differs slightly depending on how DB2 is accessed.7 DB2/UDB Enterprise stage options The DB2/UDB Enterprise (native parallel) stage should be used for reading from. and writing to a DB2 Enterprise Server Edition database with Database Partitioning Feature (DPF) • As a native.2. . communicating directly with each database node. and the target table(s) may be accessed by other users. Time and row-based commit intervals determine the transaction size. the contents of the table are unusable and the tablespace is left in a load pending state. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. transmitted.Information Integration Solutions Center of Excellence 10. forcing the DB2 Enterprise stages on each ETL server to communicate directly with their local DB2 coordinator. and the availability of new rows to other applications. the DB2 Load method places an exclusive lock on the entire DB2 tablespace into which it loads the data and no other tables in that tablespace can be accessed by other applications until the load completes. the DB2/UDB Enterprise stage is designed for maximum performance and scalability against very large partitioned DB2 UNIX databases. The choice between these methods depends on required performance. a) The Write Method (and corresponding insert / update / upsert / delete) communicates directly with the DB2 database nodes to execute instructions in parallel. DB2 Enterprise stage should not include the Client Instance Name property. No part of this publication may be reproduced.

Table T is in tablespace TS and TS is partitioned into 3 partitions on Col1 (limits: F. All rights reserved. stored in a retrieval system. . transmitted. This information determines the number of db2read operators that the conductor builds into the score and the queries that they execute. Finally the SYSKEYS and SYSCOLUMNS tables are read using the index name to get the associated column metadata (name and type). functions within the db2read operator are used to read the DB2 SYSTABLES table to retrieve the tablespace and database name for the table. and the partition limit key value(s). The WHERE clauses which are created to read this tables are: Where Col1 < ‘F’ or (Col1 = ‘F’ and (Col2 < 10 or Col2 = 10)) Where (Col1 > ‘F’ and Col1 < ‘P’) or (Col1 = ‘F’ and Col2 > 10) or (Col1 = ‘P’ and (Col2 < 20 or Col2 = 20)) Where Col1 > ‘T’ or (Col1 = ‘T’ and Col2 > 40) The method that DataStage/USS Edition uses to write to DB2 UDB on z/OS works differently than the read process.Information Integration Solutions Center of Excellence When accessing a DB2 table using the Table read method. Parallel Framework Red Book: Data Flow Job Design July 17. 30. T) and Col2 (10. and is controlled by the number of nodes in the configuration file. 40). transcribed. the number of operators do not have to match to the number of partitions. Since all write operations need to go through the DB2 coordinator node on z/OS (this is different than on non-z/OS platforms). or translated into any language in any form by any means without the written permission of IBM. as illustrated in Figure 46: Figure 46: DB2 read on DataStage/USS For example. partitioning index name(s). 2006 109 of 179 © 2006 IBM Information Integration Solutions. This is illustrated in Figure 47. P. No part of this publication may be reproduced. These values are in turn used to read the SYSTABLEPART table to retrieve the number of partitions.

An example of an in-memory Normal Lookup is shown in Figure 48. 2006 110 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 47: DB2 write on DataStage/USS On DataStage/USS Edition. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. transcribed. transmitted. Lookups work differently depending on whether the lookup is done normally (in memory) or using a sparse technique where each lookup is effectively a query to the database. . stored in a retrieval system.

Since there is no sequential file stage associated with this MVS load file. Parallel Framework Red Book: Data Flow Job Design July 17. Figure 50 illustrates the DB2 LOAD process on USS and also shows the format of the special resource statement used to define the MVS dataset used during the load operation. No part of this publication may be reproduced. transcribed. Contrast the Normal Lookup with the way a Sparse Lookup is done as shown in Figure 49. where each lookup operator is issuing an SQL to DB2 for every row it processes. In order to invoke it from a DataStage/USS job. All rights reserved. Since each of these queries must go through the DB2 coordinator node we can effectively ignore the level of parallelism specified for the table. Figure 49: DB2 Sparse Lookup on DataStage/USS Finally. nor can it be read in from a USS HFS file. transmitted. using the DB2 load utility in USS is different from non-z/OS environments. The DB2 LOAD utility is designed to run from JCL only. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. 2006 111 of 179 © 2006 IBM Information Integration Solutions. The LOAD utility has a second limitation in that data cannot be piped into it. When the conductor creates the score. .Information Integration Solutions Center of Excellence Figure 48: In-Memory Lookup on DataStage/USS Here we see that the Normal Lookup actually consists of reading the DB2 table into memory and then performing the lookup against the memory copy of the table. we call a DB2 stored procedure called DSNUTILS. we need to add a special resource statement in our configuration file to specify the MVS dataset name to use. it matches the number of db2read operators to the partitioning scheme of the table (similar to the read) and the number of lookup operators to the number of nodes in the configuration file. This requires DataStage/USS to create an MVS flat file to pass to the loader – note that this is the only non-HFS file that DS/USS can write to.

No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. . stored in a retrieval system. All rights reserved. 2006 112 of 179 © 2006 IBM Information Integration Solutions. transmitted.Information Integration Solutions Center of Excellence Figure 50: Calling DB2 Load Utility on DataStage/USS Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.

the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types .Information Integration Solutions Center of Excellence 10. A decimal with floating scale is converted to dfloat dfloat dfloat int32 decimal string[n] string[max=n] sfloat int32 sfloat int16 string[max=n] IMPORTANT: Informix data types that are not listed in the above table cannot be used in the Informix Enterprise stage.columns of the database row correspond to columns of a DS/EE record .s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n. decimal[p. If the DATETIME starts with a year component. transcribed. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. Informix Data Type CHAR(n) CHARACTER VARYING(n.s] where p is the precision and s is the scale The maximum precision is 32. the result is a timestamp field. as shown in the following table. time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month. 2006 113 of 179 © 2006 IBM Information Integration Solutions. the result is a time field. the result is a date field.r) NVARCHAR(n. All rights reserved. transmitted. DECIMAL[p. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: .r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour.3.both DS/EE and Informix support null values.rows of the database result set correspond to records of a DS/EE Data Set .3. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. . No part of this publication may be reproduced. The Informix Enterprise stage converts Informix data types to Enterprise Edition data types.2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date.3 Informix Database Guidelines 10.

4.1 ODBC Enterprise Stage Column Names For each ODBC Enterprise stage: .rows of the database result set correspond to records of a DS/EE Data Set .names are translated exactly except when the external data source column name contains a character that DataStage does not support. 2006 114 of 179 © 2006 IBM Information Integration Solutions.s] string[36] int32 int8 [0 or 1] decimal[p.4. transmitted.2 ODBC Enterprise stage Data Type Mapping ODBC data sources are not supported by the orcdbutil utility. No part of this publication may be reproduced.columns of the database row correspond to columns of a DS/EE record . .the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types .s] decimal[p. transcribed.s] int16 int8 date time[p] timestamp[p] raw[max=n] string[max=n] ustring[n] ustring[max=n] Note that the maximum size of a DataStage record is limited to 32K.both DS/EE and ODBC support null values. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. as shown in the following table: ODBC Data Type SQL_BIGINT SQL_BINARY SQL_CHAR SQL_DECIMAL SQL_DOUBLE SQL_FLOAT SQL_GUID SQL_INTEGER SQL_BIT SQL_REAL SQL_SMALLINT SQL_TINYINT SQL_TYPE_DATE SQL_TYPE_TIME SQL_TYPE_TIMESTAMP SQL_VARBINARY SQL_VARCHAR SQL_WCHAR SQL_WVARCHAR Enterprise Edition Data Type int64 raw(n) string[n] decimal[p.s] where p is the precision and s is the scale decimal[p. Parallel Framework Red Book: Data Flow Job Design July 17. Enterprise Edition will return an error and abort your job.4 ODBC Enterprise Guidelines 10. In that case. stored in a retrieval system. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.Information Integration Solutions Center of Excellence 10. If you attempt to read a record larger than 32K. two underscore characters replace the unsupported character . It is important to verify the correct ODBC to Enterprise Edition data mapping.

or translated into any language in any form by any means without the written permission of IBM. transcribed. 2006 115 of 179 © 2006 IBM Information Integration Solutions. transmitted. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. . All rights reserved. stored in a retrieval system.Information Integration Solutions Center of Excellence IMPORTANT: ODBC data types that are not listed in the above table cannot be used in the ODBC Enterprise stage.

a patch may be available through IBM IIS Support to support parallel reads through ODBC. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. Parallel reads through ODBC match the degree of parallelism in the $APT_CONFIG_FILE. . since this capability is not provided by the ODBC API.4. transmitted. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence 10. or translated into any language in any form by any means without the written permission of IBM. the ODBC Enterprise stage does not support parallel read4. All rights reserved.3 Reading ODBC Sources Unlike other native parallel database stages. 2006 116 of 179 © 2006 IBM Information Integration Solutions. it may be possible to write to a target database in parallel using the ODBC Enterprise stage. and the table configuration (row or page level lock mode if available). stored in a retrieval system. Depending on the target database. 4 On some platforms.

1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.both DS/EE and Oracle support null values. Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p. or translated into any language in any form by any means without the written permission of IBM.s] if precision (p) >=11 or scale > 0 not supported string[max=n] or ustring[max=n] a variable-length string with maximum length = n Note that the maximum size of a DataStage record is limited to 32K. Examples include: .5. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage. Enterprise Edition will return an error and abort your job.columns of the database row correspond to columns of a DS/EE record .2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. Setting the partition table option to the specified table will enable parallel extracts from an Oracle source.3 Reading from Oracle in Parallel By default.Information Integration Solutions Center of Excellence 10. two underscore characters replace the unsupported character .5.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p. .rows of the database result set correspond to records of a DS/EE Data Set .s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38. The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition. transcribed. which are not heavily typed.5. If you attempt to read a record larger than 32K.5 Oracle Database Guidelines 10. This is particularly important for Oracle databases. and will generate an error at runtime 10. It is important to note that certain types of queries cannot run in parallel. 2006 117 of 179 © 2006 IBM Information Integration Solutions.names are translated exactly except when the Oracle source column name contains a character that DataStage does not support. No part of this publication may be reproduced. the Oracle Enterprise stage reads sequentially from its source table or query. stored in a retrieval system.queries containing a GROUP BY clause that are also hash partitioned on the same field Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. In that case. transmitted.the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types .

transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. 2006 118 of 179 © 2006 IBM Information Integration Solutions. PARALLEL=FALSE) also allows loading of indexed tables without index maintenance. transcribed. rebuild). . When using this method. In this instance. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM.5. the Oracle load will be done sequentially. All rights reserved. • Setting the environment variable $APT_ORACLE_LOAD_OPTIONS to “OPTIONS (DIRECT=TRUE. The Upsert Write Method can be used to insert rows into a target Oracle table without bypassing indexes or constraints. Enterprise Edition uses the Parallel Direct Path Load method. stored in a retrieval system.4 Oracle Load Options When writing to an Oracle table (using Write Method = Load). the key column(s) must be identified using the check boxes in the column grid.Information Integration Solutions Center of Excellence - queries performing a non-collocated join (a SQL JOIN between two tables that are not stored in the same partitions with the same partitioning strategy) 10. the Oracle stage cannot write to a table that has indexes (including indexes automatically generated by Primary Key constraints) on it unless you specify the Index Mode option (maintenance. In order to automatically generate the SQL required by the Upsert method.

Information Integration Solutions Center of Excellence

10.6 Sybase Enterprise Guidelines
10.6.1 Sybase Enterprise Stage Column Names For each Sybase Enterprise stage: - rows of the database result set correspond to records of a DS/EE Data Set - columns of the database row correspond to columns of a DS/EE record - the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types - names are translated exactly except when the Sybase source column name contains a character that DataStage does not support. In that case, two underscore characters replace the unsupported character - both DS/EE and Sybase support null values, and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.6.2 Sybase Enterprise stage Data Type Mapping Sybase databases are not supported by the orcdbutil utility. It is important to verify the correct Sybase to Enterprise Edition data mapping, as shown in the following table: Sybase Data Type
BINARY(n) BIT CHAR(n) DATE DATETIME DEC[p,s] or DECIMAL[p,s] DOUBLE PRECISION or FLOAT INT or INTEGER MONEY NCHAR(n) NUMERIC[p,s] NVARCHAR(n,r) REAL SERIAL SMALLDATETIME SMALLFLOAT SMALLINT SMALLMONEY TINYINT TIME UNSIGNED INT VARBINARY(n) VARCHAR(n)

Enterprise Edition Data Type
raw(n) int8 string[n] a fixed-length string with length n date timestamp decimal[p,s] where p is the precision and s is the scale dfloat int32 decimal[15,4] ustring[n] a fixed-length string with length n - only for ASE decimal[p,s] where p is the precision and s is the scale ustring[max=n] a variable-length string with length n - only for ASE sfloat int32 timestamp sfloat int16 decimal[10,4] int8 time unit32 raw[max=n] string[max=n] a variable-length string with maximum length n

IMPORTANT: Sybase data types that are not listed in the above table cannot be used in the Sybase Enterprise stage, and will generate an error at runtime
Parallel Framework Red Book: Data Flow Job Design July 17, 2006 119 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

10.7 Teradata Database Guidelines
10.7.1 Choosing the Proper Teradata Stage Within DataStage Enterprise Edition, the following stages can be used for reading from and writing to Teradata databases in a parallel job flow: Source Teradata Stages Teradata Enterprise Teradata API Target Teradata Stages Teradata Enterprise Teradata API Teradata MultiLoad (MultiLoad option) Teradata MultiLoad (TPump option)

For maximum performance of high-volume data flows, the native parallel Teradata Enterprise stage should be used. Teradata Enterprise uses the programming interface of the Teradata utilities FastExport (reads) and FastLoad (writes), and is subject to all these utilities’ restrictions. NOTE: Unlike the FastLoad utility, the Teradata Enterprise stage supports Append mode, inserting rows into an existing target table. This is done through a shadow “terasync” table.

Teradata has a system-wide limit to the number of concurrent database utilities. Each use of the Teradata Enterprise stages counts toward this limit. 10.7.2 Source Teradata Stages Teradata Stage Stage Usage Guidelines Type
Teradata Enterprise Teradata API Native Parallel Plug-In - Reading a large number of rows in parallel - Supports OPEN and CLOSE commands - Subject to the limits of Teradata FastExport - Reading a small number of rows sequentially

Parallel Read
Yes No

Teradata Utility Limit
applies none

10.7.3 Target Teradata Stages Teradata Stage Stage Usage Guidelines Type
Teradata Enterprise Native Parallel - Writing a large number of rows in parallel - Supports OPEN and CLOSE commands - Limited to INSERT (new table) or APPEND (existing table) - Subject to the limits of Teradata FastLoad (but also supports APPEND) - Locks the target table in exclusive mode - Insert, Update, Delete, Upsert of moderate data volumes - Locks the target table(s) in exclusive mode

Parallel Write
Yes

Teradata Utility Limit
applies

Teradata MultiLoad (MultiLoad utility)

Plug-In

No

applies

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

120 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence Teradata MultiLoad (TPump utility) Plug-In - Insert, Update, Delete, Upsert of small volumes of data within a large database - Does not lock the target tables - Should not be run in parallel, because each node and use counts toward system-wide Teradata utility limit - Insert, Update, Delete, Upsert of small volumes of data - Allows concurrent writes (does not lock target) - Slower than TPump for equivalent operations

Teradata API

Plug-In

Yes

none

10.7.4 Teradata Enterprise Stage Column Names For each Teradata Enterprise stage: - rows of the database result set correspond to records of a DS/EE Data Set - columns of the database row correspond to columns of a DS/EE record - the name and data type of each database column corresponds to a DS/EE Data Set name and data type using a predefined mapping of database data types to Enterprise Edition data types - both DS/EE and Teradata support null values, and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column - DS/EE gives the same name to its columns as the Teradata column name. However, while DS/EE column names can appear in either upper or lower case, Teradata column names appear only in upper case. 10.7.5 Teradata Enterprise stage Data Type Mapping Teradata databases are not supported by the orcdbutil utility. It is important to verify the correct Teradata to Enterprise Edition data mapping, as shown in the following table: Teradata Data Type
byte(n) byteint char(n) date decimal[p,s] double precision float graphic(n) integer long varchar long vargraphic numeric(p,s) real smallint time timestamp varbyte(n) varchar(n)

Enterprise Edition Data Type
raw[n] int8 string[n] date decimal[p,s] where p is the precision and s is the scale dfloat dfloat raw[max=n] int32 string[max=n] raw[max=n] decimal[p,s] Dfloat int16 time timestamp raw[max=n] string[max=n] July 17, 2006 121 of 179

Parallel Framework Red Book: Data Flow Job Design

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Aggregates and most arithmetic operators are not allowed in the SELECT clause of a Teradata Enterprise stage.RequestedSessions=nn] where SesionsPerPlayer and RequestedSessions are optional connection parameters that are required when accessing large Teradata databases. and/or RequestedSessions should be decreased. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. Setting the SessionsPerPlayer too low on a large system can result in so many players that the job fails due to insufficient resources. In that case SessionsPerPlayer should be increased. 10. and will generate an error at runtime. where pa$$ is the example password: \’pa$$\’ 10.7. .7. Teradata Enterprise stage uses 32K buffers. and hence the number of UNIX processes and overall system resource requirements of the DataStage job. 2006 122 of 179 © 2006 IBM Information Integration Solutions.password=password[.6 Specifying Teradata Passwords with Special Characters Teradata permits passwords with special characters and symbols. SessionsPerPlayer should be set such that: RequestedSessions = (sessions per player * the number of nodes * players per node) The default value for the SessionsPerPlayer suboption is 2.7. transcribed. the DB Options property specifies the connection string and connection properties in the form: user=username.7 Teradata Enterprise Settings Within the Teradata Enterprise stage. By default. No part of this publication may be reproduced.8 Improving Teradata Enterprise Performance Setting the environment variable $APT_TERA_64K_BUFFERS may significantly improve performance of Teradata Enterprise connections depending on network configuration. 10.SessionsPerPlayer=nn][.Information Integration Solutions Center of Excellence vargraphic(n) raw[max=n] IMPORTANT: Teradata data types that are not listed in the above table cannot be used in the Teradata Enterprise stage. Parallel Framework Red Book: Data Flow Job Design July 17. RequestedSessions equals the maximum number of available sessions on the Teradata instance. Indirectly. The SessionsPerPlayer option determines the number of connections each DataStage EE player opens to Teradata. this determines the number of DataStage players. (Note that 64K buffers must be enabled at the Teradata server level). By default. but this can be set to a value between 1 and the database vprocs. transmitted. All rights reserved. the password must be surrounded by an “escaped” single quote as shown. To specify a Teradata password that contains special characters.

stored in a retrieval system. transmitted.9 Teradata on USS On the USS platform the Teradata Enterprise Stage uses CLIv2 for channel-attached systems (OS/390 and z/OS). That leaves 39 possible TDP names and is different than the convention used for non-channel attached systems. you must supply the client with the Teradata Director Program (TDP) identifier. 2006 123 of 179 © 2006 IBM Information Integration Solutions. To connect to a Teradata server. the tdpid must be in the form TDPx.Information Integration Solutions Center of Excellence 10. The first three characters must be TDP. or translated into any language in any form by any means without the written permission of IBM. transcribed. . On MVS. #.7. the tdpid is the host name of the Teradata server. Parallel Framework Red Book: Data Flow Job Design July 17. A-Z (case insensitive). $. also known as the tdpid. No part of this publication may be reproduced. where x is 0-9. or @. On a network-attached system. All rights reserved.

including CPU utilization and elapsed processing time. Normally. This can be useful when determining if the actual runtime schema matches the expected job design table definitions. Forces an internal buffer operator to be placed between every operator. lowering this number will limit the number of processes that are simultaneously created when a job is started. this can identify which part of a job flow is impacting overall performance. All rights reserved. generated osh. When trying to start very large jobs on heavilyloaded servers. Prints detailed information in the job log for each operator. the job’s partitioning logic and parallel design concepts (especially within Transformer stages) should be examined. transcribed. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. $DS_PX_DEBUG 1 $APT_PM_STARTUP_CONCURRENCY 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Prints detailed information in the job log for each operator when allocating additional heap memory. This environment variable should not normally need to be set. it is important that all jobs be tested with a configuration file that has more than one node in its default node pool. 2006 124 of 179 $APT_PM_PLAYER_TIMING 1 1 FORCE $APT_PM_PLAYER_MEMORY $APT_BUFFERING_POLICY Setting $APT_BUFFERING_POLICY=FORCE is not recommended for production job runs. Set this environment variable to capture copies of the job score. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. If the job results are correct with a single-node configuration file. No part of this publication may be reproduced. Using the job monitor performance statistics. or translated into any language in any form by any means without the written permission of IBM.1 Warning on Single-Node Configuration Files Because the DS/EE configuration file can be changed at runtime.2 Debugging Environment Variables The following environment variables can be set to assist in debugging a parallel job: Environment Variable $OSH_PRINT_SCHEMAS Setting 1 Description Outputs the actual schema definitions used by the DataStage EE framework at runtime in the DataStage log. the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance. July 17.Information Integration Solutions Center of Excellence 11 Troubleshooting and Monitoring 11. and incorrect with a multi-node configuration file. and internal Enterprise Edition log messages in a directory corresponding to the job name. stored in a retrieval system. This will ensure that the jobs have been designed with proper partitioning logic. 11. . transmitted.

it may not be possible to remove all warning messages generated by the EE engine. 11. Set the environment variable $DS_PX_DEBUG to capture copies of the job score. In some instances. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG . When a fatal error occurs.  Enable the Job Monitoring Environment Variables detailed in Section 2. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. The general process for debugging a job is:  Check the Director job log for warnings. or promote a job from development into test or production. transcribed. transmitted. or translated into any language in any form by any means without the written permission of IBM. look at row counts. generated osh.Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. and to manage source or target Parallel Data Sets. and internal Enterprise Edition log messages in a directory corresponding to the job name.3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. stored in a retrieval system. the log entry is sometimes preceded by a warning condition. All rights reserved. tune.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide.  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled). This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. This will place entries in the Director log with the actual runtime schema for every link using Enterprise Edition internal data types. The default is 30 seconds. No part of this publication may be reproduced. 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. . it is possible for $OSH_PRINT_SCHEMAS to generate a log entry that is too large for DataStage Director to store or display. All fatal and warning messages should be addressed before attempting to debug. Use the Data Set Management tool (available in the Tools menu of DataStage Designer or DataStage Manager) to examine the schema.5. July 17. To capture the full schema output in these cases.    NOTE: For large jobs. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. These may indicate an underlying logic problem or unexpected data type conversion. But all warnings should be examined and understood.

stored in a retrieval system. the generated OSH tab will appear in the Job Properties dialog box: Parallel Framework Red Book: Data Flow Job Design July 17. jobs are compiled into OSH (Orchestrate SHell) scripts that are used to execute the given job design at runtime. so if the file has any binary columns.Information Integration Solutions Center of Excellence  For flat (sequential) sources and targets: o To display the actual contents of any file in hexadecimal and ASCII (including embedded control characters or ASCII NULLs). use the UNIX command od –xc –Ax o To display the number of lines and characters in a specified ASCII text file. It is also not useful for files of non-delimited fixed-length record format. 2006 126 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. 11. NOTE: The wc command counts UNIX line delimiters. this count may be incorrect. No part of this publication may be reproduced. transcribed. and to understand internally what is running. transmitted. it must be enabled for a given project within the Administrator client: Figure 51: Generated OSH Administrator option Once this option has been enabled for a given project. It is useful to examine the generated OSH for debugging purposes.4 Viewing the Generated OSH Within Designer. use the UNIX command wc –lc [filename] Dividing the total number of characters number of lines provides an audit to ensure all rows are same length. . To enable viewing of generated OSH. All rights reserved.

where data is repartitioned and how it is repartitioned. Also available is some information about where data may be buffered. A score dump includes a variety of information about a flow. they have been optimized into the same process. the first task is to examine the score dump which is generated when you set APT_DUMP_SCORE=TRUE in your environment. . All rights reserved. or translated into any language in any form by any means without the written permission of IBM. if any. and exactly which nodes each operator runs on. 2006 127 of 179 © 2006 IBM Information Integration Solutions. ##I TFSC 004000 14:51:50(000) <main_program> This step has 1 dataset: ds0: {op0[1p] (sequential generator) eOther(APT_HashPartitioner { key={ value=a } })->eCollectAny op1[2p] (parallel APT_CombinedOperatorController:tsort)} It has 2 operators: op0[1p] {(sequential generator) on nodes ( Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. It shows three stages: Generator. what degree of parallelism each operator runs with. including how composite operators and shared containers break down. The following score dump shows a flow with a single Data Set.5 Interpreting the Parallel Job Score When attempting to understand an Enterprise Edition flow. The job runs 3 processes on 2 nodes. transcribed. have been inserted by EE. that is. No part of this publication may be reproduced. which has a hash partitioner that partitions on key field a. All stages in this flow are running on one node.Information Integration Solutions Center of Excellence Figure 52: Generated OSH in Designer Job Properties 11. Sort (tsort) and Peek. stored in a retrieval system. The Peek and Sort stages are combined. which operators.

torrent.com[op1. All rights reserved.4. Parallel Framework Red Book: Data Flow Job Design July 17.com[op0. transmitted. there are three areas to investigate: • Are there sequential stages? • Is needless repartitioning occurring? • In a cluster or Grid.Information Integration Solutions Center of Excellence lemond.com[op1. are the computation-intensive stages shared evenly across all nodes? More details on interpreting the parallel job score can be found in 12.p0] )} op1[2p] {(parallel APT_CombinedOperatorController: (tsort) (peek) )on nodes ( lemond. . No part of this publication may be reproduced. 2006 128 of 179 © 2006 IBM Information Integration Solutions.p0] lemond. or translated into any language in any form by any means without the written permission of IBM.2 Understanding the Parallel Job Score. stored in a retrieval system.torrent. transcribed.p1] )} In a score dump.torrent.

transcribed. 12. Every unused column requires additional memory which can impact performance (it also makes each transfer of a record from one stage to the next more expensive). . This section outlines performance-related tips that can be followed when building a parallel data flow using DataStage Enterprise Edition. proceeds through integration and volume testing. Parallel Framework Red Book: Data Flow Job Design July 17. • Parallel Data Sets retain data partitioning and sort order. • Lookup File Sets can be used to store reference data used in subsequent jobs. a) Use Parallel Data Sets to land intermediate result between parallel jobs. File Sets facilitate parallel I/O at the expense of exporting to a specified file format.Information Integration Solutions Center of Excellence 12 Performance Tuning Job Designs The ability to process large volumes of data in a short period of time depends on all aspects of the flow and environment being optimized for maximum throughput and performance. If you need to share information with external applications. and continues throughout an application’s production lifecycle. They maintain reference data in DS/EE internal format. use a select list to read needed columns instead of the entire table (if possible) • Be alert when using runtime column propagation (“RCP”) – it may be necessary to disable RCP for a particular stage to ensure that columns are actually removed using that stage’s Output Mapping. • Data Sets can only be read by other DS/EE parallel jobs (or the orchadmin command line utility). This section provides tips for designing a job for optimal performance. or translated into any language in any form by any means without the written permission of IBM. Performance tuning and optimization is an iterative process that begins at job design and unit tests. There are no utilities for examining data within a Lookup File Set. and for optimizing the performance of a given data flow using various settings and features within DataStage Enterprise Edition. stored in a retrieval system. Lookup File Sets can only be used on reference links to a Lookup stage. in the DS/EE native internal format. transmitted.1 How to Design a Job for Optimal Performance Overall job design can be the most significant factor in data flow performance. facilitating end-to-end parallelism across job boundaries. b) Remove unneeded columns as early as possible within the data flow. No part of this publication may be reproduced. • When reading from database sources. All rights reserved. However. pre-indexed. 2006 129 of 179 © 2006 IBM Information Integration Solutions.

transcribed. Modify) may be more appropriate. transmitted. at the expense of ease of development (and more skilled developer requirements). Note. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. For example:  Varchar columns of a large (for example. No part of this publication may be reproduced. use orchdbutil to ensure that the designtime metadata matches the actual runtime metadata (especially with Oracle databases). • When possible. o When working with database sources and targets. or when complex reusable logic is required. Because BuildOps are built in C++. When business rules dictate otherwise and the data volume is large and sufficiently skewed. especially in large-volume data flows. • There are limited scenarios when the memory overhead of handling large Varchar columns would dictate the use of unbounded strings. f) Minimize the number of partitioners in a job. other stages (for example. ensure data is as close to evenly distributed as possible. if possible. Copy. Parallel Framework Red Book: Data Flow Job Design July 17. 32K) maximum length that are rarely populated  Varchar columns of a large maximum length with highly varying data sizes d) Avoid type conversions. Use BuildOps only when existing Transformers do not meet performance requirements. however. External user-defined functions can expand the capabilities of the parallel Transformer. renaming and removing columns. repartition to a more balanced distribution as soon as possible to improve performance of downstream stages.  Enable $OSH_PRINT_SCHEMAS to verify runtime schema matches job design column definitions o Verify that the data type of defined Transformer stage variables matches the expected result type e) Minimize the number of Transformers. stored in a retrieval system. keys. that unless dynamic (parameterized) conditions are required. For data type conversions. Unbounded strings (Varchar’s without a maximum length) can have a significant negative performance impact on a job flow. • • Avoid using the BASIC Transformer. there is greater control over the efficiency of code. a Transformer is always faster than a Filter or Switch stage. . 2006 130 of 179 © 2006 IBM Information Integration Solutions. and to simply re-sort on a differing set of secondary/tertiary/etc.Information Integration Solutions Center of Excellence c) Always specify a maximum length for Varchar columns. It is usually possible to choose a smaller partition-key set.

partitioning. your results should be tested in your environment.2 Understanding Operator Combination At runtime. Combined operators are generally adjacent to each other in a data flow. As with any example. Parallel Framework Red Book: Data Flow Job Design July 17. try to maintain this sorting. o When writing to parallel Data Sets. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be used to set (in units of MB) the size of the RAM buffer for all sorts in a job. 5 One exception to this guideline is when operator combination generates too few processes to keep the processors busy. If two processes are interdependent (one processes the other’s output) and they are both CPU-bound or I/O-bound. Use SAME partitioning carefully. disabling operator combination allows CPU activity to be spread across multiple processors instead of being constrained to a single processor. The default setting is 20MB per partition. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence • • Know your data. DataStage Enterprise Edition analyzes a given job design and uses the parallel configuration file to build a job score which defines the processes and connection topology (Data Sets) between them used to execute the job logic. there is nothing to be gained from pipeline partitioning5. Enterprise Edition attempts to reduce the number of processes by combining the logic from 2 or more stages (operators) into a single process (per partition). stored in a retrieval system. transmitted. g) Minimize and combine use of Sorts where possible o It is frequently possible to arrange the order of business logic within a job flow to leverage the same sort order. sort order and partitioning are preserved. by using “SAME” partitioning. Choose hash key columns that generate sufficient unique key combinations (while satisfying business requirements). o Performance of individual sorts can be improved by increasing the memory usage per partition using the “Restrict Memory Usage (MB)” option of the standalone Sort stage. The purpose behind operator combination is to reduce the overhead associated with an increased process count. When composing the score. not explicitly stated in the sort keys) row order. specifying the “don’t sort. 12. 2006 131 of 179 © 2006 IBM Information Integration Solutions. and groupings. Remember that SAME maintains the degree of parallelism of the upstream operator.  If data has already been partitioned and sorted on a set of key columns. All rights reserved.e. previously sorted” option for those key columns in the Sort stage will reduce the cost of sorting and take greater advantage of pipeline parallelism. In these configurations. transcribed. In addition. When reading from these Data Sets. even those that have the “Restrict Memory Usage” option set. if possible. and should only be used if there is a need to maintain an implied (i. . o The stable sort option is much more expensive than non-stable sorts.

) In addition.Information Integration Solutions Center of Excellence However. the assumptions used by the Enterprise Edition optimizer to determine which stages can be combined may not always be the most efficient. All rights reserved. There are 2 ways to affect operator combination: o The environment variable APT_DISABLE_COMBINATION. stored in a retrieval system. as shown in the following illustration: Figure 53: CPU-bound combined process in Job Monitor Parallel Framework Red Book: Data Flow Job Design July 17. However. No part of this publication may be reproduced. transmitted. o Within Designer. (For information on interpreting a job score dump. When deciding which operators to include in a particular combined operator (a. it is best to let DSEE decide what to combine and what to leave uncombined.k.it will include all operators that meet the following rules: o Must be contiguous o Must be the same degree of parallelism o Must be ‘Combinable’. tuning combination might yield additional performance benefits.2 Understanding the Parallel Job Score” in this document. It is for this reason that combination can be enabled or disabled on a per-stage basis.a. if the “%CPU” column is displayed in a Job Monitor window in Director. disables ALL combination in the entire data flow. transcribed. 2006 132 of 179 © 2006 IBM Information Integration Solutions.0 versions of DS/EE.4. combination can be set on a per-stage basis (on the Stage/Advanced tab) The job score identifies what components are combined. when other performance tuning measures have been applied and still greater performance is needed. . Enterprise Edition is ‘greedy’ . Combined Operator Controller). this is only recommended on pre 7. or translated into any language in any form by any means without the written permission of IBM. or globally. here is a partial list of non-combinable operators:  Join  Aggregator  Remove Duplicates  Merge  BufferOp  Funnel  DB2 Enterprise Stage  Oracle Enterprise Stage  ODBC Enterprise Stage  BuildOps In general. combined stages are indicated by parenthesis surrounding the % CPU. see 12.

There are many factors that can reduce the number of processes generated at runtime:  Use a single-node configuration file  Remove ALL partitioners and collectors (especially when using a single-node configuration file)  Enable runtime column propagation on Copy stages with only one input and one output  Minimize join structures (any stage with more than one input. if you have several transformers and database operators combined with an output Sequential File. it is good to separate I/O heavy operators (Sequential File. in the above job design. the I/O-intensive FileSet is combined with a CPU-intensive Transformer. In fact. or translated into any language in any form by any means without the written permission of IBM. as shown in this subsequent Job Monitor for the same job: Figure 54: Throughput in Job Monitor after disabling combination 12. etc. Change Capture.3 Minimizing Runtime Processes and Resource Requirements The architecture of Enterprise Edition is well suited for processing massive volumes of data in parallel across available resources. stored in a retrieval system. transmitted. DS/EE executes a given job across the resources defined in a the specified configuration file. for example:  Batch jobs that process a small volume of data  Real-time jobs that process data in small message units  Environments running a large number of jobs simultaneously on the same server(s) In these instances. Lookup. and improves performance.). Disabling combination with the Transformer enables pipeline partitioning. This will prevent IO requests from waiting on CPU to become available and viceversa. Merge. such as Join. There are times. No part of this publication may be reproduced. however. For example. transcribed. 2006 133 of 179 © 2006 IBM Information Integration Solutions. when it is appropriate to minimize the resource requirements for a given scenario. Funnel) Parallel Framework Red Book: Data Flow Job Design July 17. it might be a good idea to set the sequential file to be non-combinable. Full Sorts. Toward that end. All rights reserved. .) from CPU-heavy operators (Transformer. in general.Information Integration Solutions Center of Excellence Choosing which operators to disallow combination for is as much art as science. etc. However. a single-node configuration file is often appropriate to minimize job startup time and resource requirements without significantly impacting overall performance.

It usually only comes into play when the size of a single record exceeds the default size of the transport block. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MAX_TRANSPORT_BLOCK_SIZE APT_MAX_TRANSPORT_BLOCK_SIZE July 17. Cannot be less than 8192. Funnel. Remove Duplicates. Oracle Enterprise. Aggregator. If necessary.4. 2006 134 of 179   Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. DB2 Enterprise. BuildOps. The default value is 8192. (Buffering is discussed in more detail in the following section. with a valid value range for between 8192 and 1048576. Merge. The first block will be used by the upstream/producer stage to output data it is done with. APT_MIN_TRANSPORT_BLOCK_SIZE o Specifies the minimum allowable block size for transferring data between players. transcribed. BufferOp Selectively (being careful to to avoid deadlocks) disable buffering. Default is 8192. No part of this publication may be reproduced. transmitted. so setting this value too high will result in a large amount of memory consumption. . cannot be greater than 1048576. ODBC Enterprise. They are actually transferred in blocks (just like on old magnetic tapes) called “Transport Blocks”. Once the upstream block is full and the downstream block is empty. This type of buffering (or ‘Record Blocking’) is rarely tuned. or translated into any language in any form by any means without the written permission of IBM. the value provided by a user for this variable is rounded up to the operating system's nearest page size. Remember. Each pair of operators that have a producer/consumer relationship will share at least 2 of these blocks. stored in a retrieval system. then setting APT_DEFAULT_TRANSPORT_BLOCK_SIZE to a multiple of (or equal to) the record size will resolve the problem.1 Inter-Operator Transport Buffering Though it may look like it from the performance statistics and documentation might discuss ‘record streaming’. 12. The behavior of these transport blocks is determined by these environment variables:  APT_DEFAULT_TRANSPORT_BLOCK_SIZE o Specifies the default block size for transferring data between players.) 12. The second block will be used by the downstream/consumer stage to obtain data that is ready for the next processing step. strictly speaking. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. there are 2 of these transport blocks for each partition of each link.Information Integration Solutions Center of Excellence   Minimize non-combinable stages (as outlined in the previous section) such as Join.4 Understanding Buffering There are two types of buffering in Enterprise Edition: ‘inter-operator transport’ and ‘deadlock prevention’. the blocks are swapped and the process begins again. All rights reserved. records do not stream from one stage to another.

transmitted. Cannot be less than 8192.2 Deadlock Prevention Buffering The other type of buffering. No part of this publication may be reproduced. . the framework calculates the block size for transferring data between players according to this algorithm: if (recordSize * APT_LATENCY_COEFFICIENT < APT_MIN_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MIN_TRANSPORT_BLOCK_SIZE else if (recordSize * APT_LATENCY_COEFFICIENT > APT_MAX_TRANSPORT_BLOCK_SIZE) then blockSize = APT_MAX_TRANSPORT_BLOCK_SIZE else blockSize = recordSize * APT_LATENCY_COEFFICIENT APT_LATENCY_COEFFICIENT o Specifies the number of records to be written to each transport block APT_SHARED_MEMORY_BUFFERS o Specifies the number of Transport Blocks between a pair of operators.4. 12. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE. transcribed. APT_LATENCY_COEFFICIENT. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. Here is an example job fragment: Figure 55: Fork-Join example Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. All rights reserved.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players. 2006 135 of 179 © 2006 IBM Information Integration Solutions. “Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. cannot be greater than 1048576. Default is 1048576. or translated into any language in any form by any means without the written permission of IBM.

d eue Qu ad Re Transformer Waiting to write to Aggregator1 Qu eue Re d ad Join Waiting to read from Aggregator2 July 17. not necessarily a Join stage) by Enterprise Edition during job startup. All rights reserved. that “Fork-Join” is a graphical description. which go into an Inner Join. and Join is waiting to read from Aggregator2. No part of this publication may be reproduced. stored in a retrieval system. To understand deadlock-prevention buffering. Aggregator2 is waiting to read from the Transformer.bringing processing to a halt (though the job does not stop running. Join is waiting on Aggregator2. It is placed on all inputs to a join structure (again. the job would deadlock . and Aggregator2 is waiting on Transformer.Information Integration Solutions Center of Excellence In this example. this scenario would create a circular dependency where Transformer is waiting on Aggregator1. it doesn’t necessarily have to involve a Join stage. Aggregator1 is waiting on Join. instead of data flow. BufferOp is always ready to read or write and will not allow a read/write request to be queued.) Without deadlock buffering. transmitted. or translated into any language in any form by any means without the written permission of IBM. Without deadlock buffering. it is important to understand the operation of a parallel pipeline. the Transformer creates a fork with 2 parallel Aggregators. transcribed. Aggregator1 is waiting to write to the Join. Note however. Imagine that the Transformer is waiting to write to Aggregator1. there is a specialized operator called BufferOp. Like this: Aggregator 1 Waiting to Write to Join d e ue Qu ite Wr Qu eu Wr ed ite Aggregator2 Waiting to read from Transformer (Here the arrows represent dependency direction. 2006 136 of 179 . it would eventually time out). To guarantee that this problem never happens in Enterprise Edition. So the above job structure would be altered by the DS/EE engine to look like this: Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.

transmitted. BufferOps will also be placed on the input partitions to any sequential stage that is fed by a parallel stage. Default 1 MB. TIP: For very wide rows. APT_BUFFER_FREE_RUN o Maximum capacity of the buffer operator before it starts to offer resistance to incoming flow. Values greater than 1 indicate that the buffer July 17.) Since BufferOp is always ready to read or write. as these same types of circular dependencies can result from partition-wise Fork-Joins.Information Integration Solutions Center of Excellence BufferOp1 Aggregator 1 Transformer Join Aggregator2 BufferOp 2 (Here the arrows now represent data-flow. BufferOps will allocate 3MB of memory each (remember that this is per operator. No part of this publication may be reproduced. Join cannot be ‘stuck’ waiting to read from either of its inputs. 2006 137 of 179  Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. May not exceed 2/3 of APT_BUFFER_MAXIMUM_MEMORY. “The Parallel Engine Configuration File” of the DataStage Manager guide). stored in a retrieval system. thus breaking the circular dependency and guaranteeing no deadlock will occur. The behavior of deadlock-prevention BufferOps can be tuned through these environment variables:  APT_BUFFER_DISK_WRITE_INCREMENT o Controls the “blocksize” written to disk as the memory buffer fills. By default. not dependency. as a nonnegative (proper or improper) fraction of APT_BUFFER_MAXIMUM_MEMORY. . or translated into any language in any form by any means without the written permission of IBM. it may be necessary to increase the default buffer size (APT_BUFFER_MAXIMUM_MEMORY) to hold more rows in memory. transcribed. per partition). When that is full (because the upstream operator is still writing but the downstream operator isn’t ready to accept that data yet) it will begin to flush data to the scratchdisk resources specified in the configuration file (detailed in Chapter 11. All rights reserved.

queue bound.  APT_BUFFER_MAXIMUM_MEMORY o Maximum memory consumed by each buffer operator for data storage. the default buffering policy is AUTOMATIC_BUFFERING. .Information Integration Solutions Center of Excellence operator will free run (up to a point) even when it has to write data to disk. BufferOps also have the effect of “smoothing out” production/consumption spikes. Additionally. buffer size. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. transcribed. use of this setting is only recommend for advanced users! FORCE_BUFFERING can be used to reveal bottlenecks in a job design during development and performance tuning. Valid settings are:  AUTOMATIC_BUFFERING: buffer as necessary to prevent dataflow deadlocks  FORCE_BUFFERING: buffer all virtual Data Sets  NO_BUFFERING: inhibit buffering on all virtual Data Sets  WARNING: Inappropriately specifying NO_BUFFERING can cause dataflow deadlock during job execution. buffer free run. transmitted. but will almost certainly degrade performance and therefore shouldn’t be used in production job runs. Default is 3 MB. APT_BUFFERING_POLICY o Specifies the buffering policy for the entire job. No part of this publication may be reproduced. When it is not defined or defined to be the null string. the buffer mode. 2006 138 of 179 © 2006 IBM Information Integration Solutions. as shown in the illustration below: Aside from ensuring that no dead-lock occurs. and write increment can be set on a per-stage basis from the Input/ Advanced tab of the stage properties. All rights reserved. This allows the job to run at the highest rate possible even when a Parallel Framework Red Book: Data Flow Job Design July 17.

transmitted. When attempting to address these mismatches in production/consumption. NFS mount points for data storage. it is best to tune the buffers on a per-stage basis. etc. Choosing which stages to tune buffering for and which to leave alone is as much art as science. while changing only 1 thing at a time). No part of this publication may be reproduced. you cannot determine that any one stage is waiting on any other stage. transcribed. then setting the buffering policy to “FORCE_BUFFERING” will cause buffering to occur everywhere. you may be able identify points in the data flow where a downstream stage is waiting on an upstream stage to produce data.this is called “pushback”. As implied above. 2006 139 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17.Information Integration Solutions Center of Excellence downstream stage is ready for data at different times than when its upstream stage is ready to produce that data. or translated into any language in any form by any means without the written permission of IBM. Stages upstream/downstream from high-latency stages (such as remote databases.) are a good place to start. when a buffer has consumed its RAM. as some other stage far downstream could be responsible for cascading pushback all the way upstream to the place you are seeing the bottleneck. . All rights reserved. instead of globally through environment variable settings. if you do not have force buffering set and APT_BUFFER_FREE_RUN set to at least ~1000. Each place may offer an opportunity for buffer tuning. and should be considered among the last resorts for performance tuning. it will ask the upstream stage to “slow down” . Because of this. If that doesn’t yield enough of a performance boost (remember to test iteratively. By using the performance statistics in conjunction with this buffering.

5. or translated into any language in any form by any means without the written permission of IBM. along with cross-references for more detail. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. 1. All rights reserved. An example directory naming structure is given in Section 2. See Section 2. An example DataStage naming structure is given in Section 2.  Naming conventions. 2006 140 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. DataStage Template jobs should be created with:  standard parameters (for example. Development Guidelines Modular development techniques should be used to maximize re-use of DataStage jobs and components.1:Directory Structures. etc) and directories where files are stored. database login properties…)  environment variables and their default settings (as outlined in Section 2. This can also be used for integration with source code control systems. as outlined in Section 3:Development Guidelines:  Job parameterization allows a single job design to process similar logic instead of creating multiple copies of the same job. stage names. Parallel Framework Red Book: Data Flow Job Design July 17. and links.5: Job Parameters. and so forth.3: Documentation and Annotation. transcribed.Information Integration Solutions Center of Excellence Appendix A: Standard Practices Summary This Appendix summarizes Standard Practices recommendations outlined in this document.  Where possible. file names. create re-usable components such as parallel shared containers to encapsulate frequently-used logic.1 Environment Variable Settings)  annotation blocks Job Parameters should always be used for file paths. source and target file paths. especially for DataStage Project categories. .  A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN.2: Naming Conventions. No part of this publication may be reproduced. transmitted. Standards It is important to establish and follow consistent standards in:  Directory structures for install and application support directories. The scope of a parameter is discussed further in Section 3. database login settings. password. these standard parameters and settings should be made part of a Designer Job Template.4: Working with Source Code Control Systems. To ease re-use. using the Manager DSX export capability. intermediate work files. See Section 2. as well as Annotation fields.  Create a standard directory structure outside of the DataStage project directory for source and target files. All DataStage jobs should be documented with Short Description field. It is the DataStage developer’s responsibility to make personal backups of their work on their local workstation. 2. user.

stored in a retrieval system.Information Integration Solutions Center of Excellence Parallel Shared Containers should be used to encapsulate frequently-used logic. transcribed. Standardized Error Handling routines should be followed to capture errors and rejects. 4. No part of this publication may be reproduced. Server Shared Containers) within a parallel job. . 5. while minimizing overhead.  Always use parallel Data Sets for intermediate storage between jobs. Objective 2: The partition method must match the business requirements and stage functional requirements.   DataStage Data Types Be aware of the mapping between DataStage (SQL) data types and the internal DS/EE data types. or translated into any language in any form by any means without the written permission of IBM. 3. Leverage default type conversions using the Copy stage or across the Output mapping tab of other stages. Further details are provided in Section 3. All rights reserved. and to facilitate default type conversions.  Use the parallel Transformer stage (not the BASIC Transformer) instead of the Filter or Switch stages.7:Error and Reject Record Handling.8: Component Usage. BASIC Routines are appropriate only for job control sequences. the following objectives help to form a methodology for assigning partitioning: Objective 1: Choose a partitioning method that gives close to an equal number of rows in each partition. as outlined in Section 4:DataStage Data Types. minimizing overall run time. 2006 141 of 179 © 2006 IBM Information Integration Solutions. transmitted.  Use the Copy stage as a placeholder for iterative design.  Use BuildOp stages only when logic cannot be implemented in the parallel Transformer. Component Usage As discussed in Section 3. using RCP to maximize re-use. the following guidelines should be followed when constructing parallel jobs in DS/EE:  Never use Server Edition components (BASIC Transformer. Partitioning Data Given the numerous options for keyless and keyed partitioning. This ensures that the processing workload is evenly balanced. assigning related records to the same partition if required Any stage that processes groups of related records (generally using one or more key columns) must be partitioned using a keyed partition method. Parallel Framework Red Book: Data Flow Job Design July 17.

Merge. Collecting Data Given the options for collecting data into a sequential stream. 6. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. All rights reserved. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. This is particularly useful if downstream jobs are run with the same degree of parallelism (configuration file) and require the same partition and sort order. stored in a retrieval system. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. Change Capture. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. use Round Robin partitioning to redistribute data equally across all partitions o Especially useful if the input Data Set is highly skewed or sequential d) Use Same partitioning to optimize end-to-end partitioning and to minimize repartitioning o Being mindful that Same partitioning retains the degree of parallelism of the upstream stage o Within a flow. and Sort stages. or translated into any language in any form by any means without the written permission of IBM. but is not limited to: Aggregator. 2006 142 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence This includes. Join. . Objective 3: Unless partition distribution is highly skewed. Using the above objectives as a guide. persistent Data Sets can be used to retain the partitioning and sort order. Remove Duplicates. minimize repartitioning. Objective 4: Partition method should not be overly complex The simplest method that meets the above objectives will generally be the most efficient and yield the best performance. transmitted. Note that in satisfying the requirements of this second objective. It may also be necessary for Transformers and BuildOps that process groups of related records. No part of this publication may be reproduced. the following methodology can be applied: a) Start with Auto partitioning (the default) b) Specify Hash partitioning for stages that require groups of related records o Specify only the key column(s) that are necessary for correct grouping as long as the number of unique values is sufficient o Use Modulus partitioning if the grouping is on a single integer key column o Use Range partitioning if the data is highly skewed and the key column values and distribution do not change significantly over time (Range Map can be reused) c) If grouping is not required. use Auto partitioning (the default) Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. Change Apply. Across jobs.

The Lookup stage is most appropriate when reference data is small enough to fit into available memory. stored in a retrieval system.1. o Always place a reject link on a parallel Transformer to capture / audit possible rejects.  Stage-Specific Guidelines As discussed in Section 8. .  Parallel Framework Red Book: Data Flow Job Design July 17. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting.1: Transformer NULL Handling and Reject Link. transcribed. or translated into any language in any form by any means without the written permission of IBM. No part of this publication may be reproduced. Output Statistics  Always specify “DataStage” Sort Utility for standalone Sort stages  Use the “Sort Key Mode=Don’t Sort (Previously Sorted)” to resort a sub-grouping of a previously-sorted input Data Set e) Be aware of automatically-inserted sorts  Set $APT_SORT_INSERTION_CHECK_ONLY to verify but not establish required sort order f) Minimize the use of sorts within a job flow g) To generate a single. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. the Ordered collector may be more efficient c) Use a Round Robin collector to reconstruct rows in input order for round-robin partitioned input Data Sets. All rights reserved. 2006 143 of 179 © 2006 IBM Information Integration Solutions. Create Key Change Column. Create Cluster Key Change Column.Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel. transmitted. the following methodology should be applied when sorting in a DataStage Enterprise Edition data flow: a) Start with a link sort b) Specify only necessary key column(s) c) Don’t use Stable Sort unless needed d) Use a stand-alone Sort stage instead of a Link sort for options that not available on a Link sort:  Sort Key Mode. Merge. 7. sequential ordered result set use a parallel Sort and a Sort Merge collector 8. as long as the Data Set has not been repartitioned or reduced. use Sort Merge collector to produce a single. If the Data Sets are larger than available memory resources. use the Join or Merge stage. See Section 9.1: Lookup vs. precautions must be taken when using expressions or derivations on nullable columns within the parallel Transformer: o Always convert nullable columns to in-band values before using them in an expression or derivation. Join vs. Sorting Using the rules and behavior outlined in Section 6: Sorting.

 Database Stage Guidelines Where possible. as discussed in Section 10. Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. use a SQL where clause to limit the number of rows sent to a DataStage job. (See Section 9. Check the Director log for warnings.Information Integration Solutions Center of Excellence    Limit the use of database Sparse Lookups to scenarios where the number of input rows is significantly smaller (for example 1:100 or more) than the number of reference rows. Even if the source data is not nullable. A Sort method Aggregator should be used when the number of distinct key values is large or unknown. or translated into any language in any form by any means without the written permission of IBM. 9. If possible. When using Oracle. Care must be taken to observe the data type mappings documented in Section 10: Database Stage Guidelines when designing a parallel job with DS/EE.   Troubleshooting and Monitoring Always test DS/EE jobs with a parallel configuration file ($APT_CONFIG_FILE) that has two or more nodes in its default pool. which may indicate an underlying problem or data type conversion issue. use the native parallel database stages for maximum performance and scalability. Use Hash method Aggregators only when the number of distinct key column values is small.1. DB2. the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records.1: Database stage types: Native Parallel Database Stages DB2/UDB Enterprise Informix Enterprise ODBC Enterprise Oracle Enterprise SQL Server Enterprise Teradata Enterprise      The ODBC Enterprise stage should only be used when a native parallel stage is not available for the given source or target database. transcribed. it is best to implement business rules natively using DataStage parallel components. No part of this publication may be reproduced. or Informix databases. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. use orchdbutil to properly import design metadata. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. stored in a retrieval system. transmitted.2Capturing Unmatched Records from a Join). . 10. For maximum scalability and parallel performance. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow. or when exception processing. 2006 144 of 179 © 2006 IBM Information Integration Solutions.

2006 145 of 179 © 2006 IBM Information Integration Solutions. Set the environment variable $OSH_PRINT_SCHEMAS to capture actual runtime schema to the Director log. transmitted. No part of this publication may be reproduced. and examine the job score by following the guidelines outlined in Section 12.2 Understanding the Parallel Job Score.4. Enable $APT_DUMP_SCORE by default. All rights reserved. transcribed. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. Set $DS_PX_DEBUG if the schema record is too large to capture in a Director log entry. error and warning messages from a running DS/EE job. .Information Integration Solutions Center of Excellence    The environment variable $DS_PX_DEBUG can be used to capture all generated OSH. or translated into any language in any form by any means without the written permission of IBM.

stored in a retrieval system. transcribed.Information Integration Solutions Center of Excellence Appendix B: DataStage Naming Reference Every name should be based on a three-part concept: Subject. All rights reserved. or the function the object performs: Project Repository and Components <proj>Dev Development <proj>Test Test <proj>Prod Production BdOp<name> BuildOp XFn<name> Parallel External Function Wrap<name> Wrapper Job Names and Properties Extract Job Load Sequence Parallel Shared Container Server Shared Container Parameter Links (prefix with “lnk_”) Reference (Lookup) Reject (Lookup. transmitted. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. Class Word where the following frequently-used Class Words describe the object type. DB) Message (Sequence) Get (Shared Container) Put (Shared Container) Input Output Delete Insert Update Data Store Database Stored Procedure Table View Dimension Fact Source Target Development / Debug Stages Column Generator Head Peek Row Generator Sample Src<job> Load<job> <job>_Seq <job>Psc <job>Ssc <name>Parm Ref Rej Msg Get Put In Out Del Ins Upd DB SP Tab View Dim Fact Src Tgt CGen Head Peek RGen Smpl Tail File Stages Sequential File Complex Flat File File Set Parallel Data Set Lookup File Set External Source External Target Parallel SAS Data Set Processing Stages Aggregator Change Apply Change Capture Copy Filter Funnel Join (Inner) Join (Left Outer) Join (Right Outer) Join (Full Outer) Lookup Merge Modify Pivot Remove Duplicates SAS processing Sort Surrogate Key Generator Switch Transformer Stage Transformer (native parallel) BASIC Transformer (Server) Stage Variable Real Time Stages RTI Input RTI Output XML Input July 17. File. . Subject Modifier. 2006 Tail SF CFF FS DS LFS XSrc XTgt SASd Agg ChAp ChCp Cp Filt Funl InJn LOJn ROJn FOJn Lkp Mrg Mod Pivt RmDp SASp Srt SKey Swch Tfm BTfm SV RTIi RTIo XMLi 146 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.

2006 147 of 179 © 2006 IBM Information Integration Solutions. transcribed. . or translated into any language in any form by any means without the written permission of IBM. transmitted. stored in a retrieval system. No part of this publication may be reproduced. All rights reserved.Information Integration Solutions Center of Excellence XML Output XML Transformer Restructure Stages Column Export Column Import XMLo XMLt CExp CImp Parallel Framework Red Book: Data Flow Job Design July 17.

1. No part of this publication may be reproduced. not the actual job. As shown in the illustration below. the DS/EE job score: • • • • • Identifies degree of parallelism and node assignment(s) for each operator Details mappings between functional (stage/operator) and actual operating system processes Includes operators automatically inserted at runtime: o Buffer operators to prevent deadlocks and optimize data flow rates between stages o Sorts and Partitioners that have been automatically inserted to ensure correct results Outlines connection topology (Data Sets) between adjacent operators and/or persistent Data Sets Defines number of actual operating system processes Where possible.Information Integration Solutions Center of Excellence Appendix C: Understanding the Parallel Job Score Jobs developed in DataStage Enterprise Edition are independent of the actual hardware and degree of parallelism used to run the job. and interconnects (Data Sets) between them. Parallel Framework Red Book: Data Flow Job Design July 17. It is recommended that this setting be enabled by default at the project level. 2006 148 of 179 © 2006 IBM Information Integration Solutions. as the job score offers invaluable data for debugging and performance tuning. stored in a retrieval system. The second score entry is the actual job score. or translated into any language in any form by any means without the written permission of IBM. The first score is from the license operator. and can be ignored. the job score is output to the DataStage Director log. Enterprise Edition uses the given job design and configuration file to compose a job score which details the processes created. and the overhead to capture the score is negligible. All rights reserved. job score entries start with the phrase “main_program: This step has n datasets…” Two separate scores are written to the log for each job run. The parallel Configuration File provides a mapping at runtime between the compiled job and the actual runtime infrastructure and resources by defining logical processing nodes. transcribed. multiple operators are combined within a single operating system process to improve performance and optimize resource requirements. At runtime. Viewing the Job Score When the environment variable APT_DUMP_SCORE is set. Similar to the way a parallel database optimizer builds a query plan. transmitted. . degree of parallelism and node (server) assignments.

and the degree of parallelism per operator o Node assignment for each operator.Information Integration Solutions Center of Excellence Actual job score 2. Terminology in this section can be used to identify the type of partitioning or collecting that was used between operators. as shown in the example on the right: • Data Sets: starts with the words “main_program: This step has n datasets:” The first section details all Data Sets. there are 3 operators. for a total of 9 operating system process. • Operators: starts with the words “It has n operators:” The second section details actual operators created to execute the job flow. or translated into any language in any form by any means without the written permission of IBM. . one running sequentially. including persistent (on disk) and virtual (in memory. links between stages). All rights reserved. “node4”). In this example. In this example. there are two virtual Data Sets. 2006 149 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. This includes: o Sequential or Parallel operation. The actual node names correspond to node names in the parallel configuration file. (in this example: “node1”. transmitted. No part of this publication may be reproduced. transcribed. “node3”. two running in parallel across 4 nodes. July 17. stored in a retrieval system. Parallel Job Score Components The Enterprise Edition parallel job score is divided into two sections. “node2”.

not SAME) Parallel producer to Sequential consumer No producer or no consumer (typically. In the above example. The symbol between the partition name and collector name indicates: -> <> => #> >> > Sequential producer to Sequential consumer Sequential producer to Parallel consumer Parallel producer to Parallel consumer (SAME partitioning) Parallel producer to Parallel consumer (repartitioned. with each Data Set identified by its number (starting at zero). Data Sets are identified in the first section of the parallel job score. for persistent Data Sets) July 17. . while operators are identified by their operator number and name. All rights reserved. 3. 2006 150 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. followed by the downstream consumer. collector type with the second. transcribed. The memory used by deadlock-prevention BufferOps can be calculated based on the number of inserted BufferOps. the upstream producer is identified first. Operator 1 (op1) is running in parallel with 4 degrees of parallelism [4p]. The partition type is associated with the first term. Job Score: Data Sets The parallel pipeline architecture of DataStage Enterprise Edition passes data from upstream producers to downstream consumers through in-memory virtual data sets. with 1 degree of parallelism [1p]. the first Data Set is identified as “ds0”. transmitted. Within the Data Set definition.Information Integration Solutions Center of Excellence Note that the number of virtual Data Sets and the degree of parallelism determine the amount of memory used by the inter-operator transport buffers. corresponding to the lower section of the job score. Producer Partitioner Collector Consumer The notation between producer and consumer is used to report the type of partitioning or collecting (if any) that is applied. operator zero (op0) is running sequentially. Persistent Data Sets are identified by their Data Set name. in the example on the right. No part of this publication may be reproduced. followed by a notation to indicate the type of partitioning or collecting (if any). and the next “ds1”. Producers and consumers may be either persistent (on disk) Data Sets or parallel operators. stored in a retrieval system. For example. as illustrated in the example on the right: The degree of parallelism is identified in brackets after the operator name.

p0] o may also include combined node2[op2.p3] (APT_CombinedOperatorController) which include logic from multiple stages in a single operator o may also include framework-inserted operators such as Buffers. if the Preserve Partitioning flag has been set for a particular Data Set.Information Integration Solutions Center of Excellence Finally. as shown in the following score fragment shown on the right: ecc3673[op3.p3] )} op4[4p] {(parallel APT_CombinedOperatorController: (APT_LUTProcessImpl in Lookup_3) (APT_TransformOperatorImplV0S7_cpLookupTest1_Tran sformer_7 in Transformer_7) (PeekNull) ) on nodes ( ecc3671[op4. 4. this (inserted tsort operator {key={value=LastName}.p2] node4[op2.p3] )} specified stage name in the Designer op2[4p] {(parallel buffer(0)) canvas on nodes ( node1[op2. But internally. No part of this publication may be reproduced. For example.p1] At runtime. Lookup is a composite operator. the notation “[pp]” will appear in this section of the job score. a composite operator includes more than one function. Job Score: Operators The lower portion of the parallel job score details op0[1p] {(sequential the mapping between stages and actual processes APT_CombinedOperatorController: (Row_Generator_0) generated at runtime. It is composted of the following internal operators: . each individual component of a composite operator is represented as an individual operator in the job score.p1] • Components of the operator node3[op2. .p1] operators node3[op2. All rights reserved. a composite operator appears to be a single stage on the design canvas. (example “[4p]”) key={value=FirstName}}(0)) on nodes ( • “sequential” or “parallel” execution mode node1[op2. transcribed.p0] from zero (example “op0”) )} op1[4p] {(parallel inserted tsort operator • degree of parallelism within brackets {key={value=LastName}. Sorts o may include “composite” operators (for example. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. 2006 151 of 179 © 2006 IBM Information Integration Solutions.p2] ecc3674[op3. Lookup) Some stages are composite operators – to the DataStage developer.p0] data has been loaded ecc3672[op3. transmitted.p0] node2[op2. Since Parallel Framework Red Book: Data Flow Job Design July 17.APT_LUTProcessImpl: op3[4p] {(parallel buffer(0)) Performs actual lookup processing once reference on nodes ( ecc3671[op3. For each operator.p2] Using this information together with the output from the $APT_PM_SHOW_PIDS environment variable. includes (as illustrated in the job score fragment): key={value=FirstName}}) ) on nodes ( • operator name (opn) numbered sequentially node1[op0.p0] )} .p2] o typically correspond to the usernode4[op2.APT_LUTCreateImpl: op2[1p] {(parallel APT_LUTCreateImpl in Lookup_3) on nodes ( Reads the reference data into memory ecc3671[op2.p1] ecc3673[op4. you can evaluate the memory used by a lookup.p0] ecc3672[op4.

you can also determine the delay associated with loading the lookup structure.Delete Data Files ->eCollectAny op2[1p] (sequential delete descriptor file in delete temp.ds) on nodes ( node1[op1. In a similar way. All rights reserved. or translated into any language in any form by any means without the written permission of IBM.Delete Descriptor File ds1: {op0[1p] (sequential Row_Generator_0) -> temp. stored in a retrieval system. a persistent Data Set defined to “Overwrite” an existing Data Set of the same name main_program: This step has 2 datasets: will have multiple entries in the job score to: ds0: {op1[1p] (parallel delete data files in delete temp. No part of this publication may be reproduced. 2006 152 of 179 © 2006 IBM Information Integration Solutions.p0] )} Parallel Framework Red Book: Data Flow Job Design July 17.ds) on nodes ( node1[op2.ds) .ds)} . transmitted.p0] )} op2[1p] {(sequential delete descriptor file in delete temp.p0] )} op1[1p] {(parallel delete data files in delete temp. . transcribed.ds} It has 3 operators: op0[1p] {(sequential Row_Generator_0) on nodes ( node1[op0.Information Integration Solutions Center of Excellence the entire structure needs to be loaded before actual lookup processing can begin.

transcribed. As can APT_Transfer::getTransferBufferSize(). 2006 153 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. if you have a transfer that transfers all fields from input to output.Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. transmitted. Parallel Framework Red Book: Data Flow Job Design July 17.add (# nullable fields)/8 for null indicators . or translated into any language in any form by any means without the written permission of IBM. the method APT_Record::estimateFinalOutputSize() can give you an estimate for a given record schema.one byte per column for field alignment (worst case is 3. rounded up 8 bytes n + 4 bytes for non-NLS data 2n + 4 bytes for NLS data (internally stored as UTF-16) n bytes for non-NLS data 2n bytes for NLS data 4 bytes 8 bytes with microsecond resolution 4 bytes 8 bytes 12 bytes with microsecond resolution For the overall record width: . All rights reserved.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries. this Appendix provides a more accurate and detailed way to estimate the size of a parallel Data Set based on the internal storage requirements for each data type: Data Type Integers Small Integer Tiny Integer Big Integer Decimal Float VarChar(n) Char(n) Time Date Timestamp Size 4 bytes 2 bytes 1 byte 8 bytes (precision+1)/2. . No part of this publication may be reproduced.

In some disk array configurations. setting this variable to a value equal to the read / write size in bytes can improve performance of Sequential File import/export operations. imported string fields that exceed their maximum declared length are truncated.01 and later) [Kbytes] $APT_IMPORT_BUFFER_SIZE $APT_EXPORT_BUFFER_SIZE $APT_CONSISTENT_BUFFERIO_SIZE [bytes] Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. These variables can be used on an as-needed basis to tune the performance of a particular job flow. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. Environment Variable $APT_EXPORT_FLUSH_COUNT Sequential File Stage Environment Variables Setting [nrows] Description Specifies how frequently (in rows) that the Sequential File stage (export operator) flushes its internal buffer to disk. to assist in debugging. Default is 128 (128K). with a minimum of 8. All rights reserved. or to change the default behavior of specific DataStage Enterprise Edition stages. transmitted. a variable-length string field to a fixed length (or a fixed-length to a longer fixed-length). July 17. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Appendix E: Environment Variable Reference This Appendix summarizes the environment variables mentioned throughout this document. NOTE: The environment variable settings in this Appendix are only examples. stored in a retrieval system. 1.1. transcribed. Setting this environment variable directs DataStage to reject Sequential File records with strings longer than their declared maximum column length. No part of this publication may be reproduced. See section 4. Set values that are optimal to your environment. By default. Defines size of I/O buffer for Sequential File reads (imports) and writes (exports) respectively. . 2006 154 of 179 $APT_IMPORT_REJECT_STRING_FIELD_OVERRUN S 1 (DataStage v7. Environment Variable $APT_STRING_PADCHAR Job Design Environment Variables Setting [char] Description Overrides the default pad character of 0x0 (ASCII null) used when EE extends. but there is a small performance penalty from increased I/O. Increasing these values on heavily-loaded file servers may improve performance. Setting this value to a low number (such as 1) is useful for realtime applications.2: Default and Explicit Type Conversions 2. or pads.

All rights reserved. $APT_DB2INSTANCE_HOME $APT_DBNAME $APT_RDBMS_COMMIT_ROWS Can also be specified with the “Row Commit Interval” stage input property. transmitted. transcribed. Sequential File (import) will read ahead 500 bytes to get the next delimiter. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. $DB2DBDFT is used to find the database name. If $APT_DBNAME is not defined.Information Integration Solutions Center of Excellence $APT_DELIMITED_READ_SIZE [bytes] Specifies the number of bytes the Sequential File (import) stage reads-ahead to get the next delimiter. this will significantly improve file pattern reads. Used as a backup for specifying the DB2 installation directory (if $INSTHOME is undefined). and so on (4X) up to 100. If it is not found the importer looks ahead 4*500=2000 (1500 more) bytes. This setting should be set to a lower value when reading from streaming inputs (for example. The default is 500 bytes. Allows DataStage plug-in stages to handle DB2 databases which use the special characters # and $ in column names. but this can be set as low as 2 bytes.000 bytes. This variable controls the upper bound which is by default 100. $DS_ENABLE_RESERVED_CHAR_CONVE RT [rows] 1 Parallel Framework Red Book: Data Flow Job Design July 17. $APT_MAX_DELIMITED_READ_SIZE [bytes] $APT_IMPORT_PATTERN_USES_FILESET [set] 3. Specifies the number of records to insert between commits. For disk configurations with multiple controllers and disk. This variable is usually set in a user’s environment from .db2profile. No part of this publication may be reproduced. Environment Variable $INSTHOME DB2 Environment Variables Setting [path] [path] [database] Description Specifies the DB2 install directory. use this variable instead of APT_DELIMITED_READ_SIZE. When more than 500 bytes read-ahead is desired. When this environment variable is set (present in the environment) file pattern reads are done in parallel by dynamically building a File Set header based on the list of files that match the given expression. . 2006 155 of 179 © 2006 IBM Information Integration Solutions. Specifies the name of the DB2 database for DB2/UDB Enterprise stages if the “Use Database Environment Variable” option is True.000 bytes. socket or FIFO) to avoid blocking. By default. The default value is 2000 per partition.

this is set to OPTIONS(DIRECT=TRUE. whichever comes first. Specifies the Oracle service name. Specifies the path to the Informix sqlhosts file. stored in a retrieval system. By default. commits are made every 2 seconds or 5000 rows per partition. transcribed. . Informix Environment Variables Environment Variable $INFORMIXDIR $INFORMIXSQLHOSTS $INFORMIXSERVER $APT_COMMIT_INTERVAL Setting [path] [filepath] [name] [rows] Description Specifies the Informix install directory. Commits are made whenever the time interval period has passed or the row interval is reached. Specifies the name of the Informix server matching an entry in the sqlhosts file. corresponding to a TNSNAMES entry.01 and later) $APT_ORA_IGNORE_CONFIG_FILE_PARALLELIS M 1 $APT_ORA_WRITE_FILES [filepath] Specifies a field delimiter for target Oracle stages using the Load method. The default is 10000 per partiton. When set. When set. 5. Normally set in a user’s environment by Oracle scripts. No part of this publication may be reproduced. Specifies Oracle SQL*Loader options used in a target Oracle stage with Load method. the output of a Target Oracle stage with Load method is written to files instead of invoking the Oracle SQL*Loader. transmitted. July 17. PARALLEL=TRUE) $ORACLE_SID $APT_ORAUPSERT_COMMIT_ROW_INTERVAL $APT_ORAUPSERT_COMMIT_TIME_INTERVAL $APT_ORACLE_LOAD_OPTIONS [SQL* Loader options] [char] $APT_ORACLE_LOAD_DELIMITED (DataStage 7. Specifies the commit interval in rows for Informix HPL Loads. a target Oracle stage with Load method will limit the number of players to the number of datafiles in the table’s tablespace. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence 4. Environment Variable $ORACLE_HOME Oracle Environment Variables Setting [path] [sid] [num] [seconds] Description Specifies installation directory for current Oracle instance. By default. All rights reserved. Useful in debugging Oracle SQL*Loader issues. Setting this variable makes it possible to load fields with trailing or leading blank characters. These two environment variables work together to specify how often target rows are committed for target Oracle stages with Upsert method. 2006 156 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The filepath specified by this environment variable specifies the file with the SQL*Loader commands.

Information Integration Solutions Center of Excellence
$DS_ENABLE_RESERVED_CHAR_CONVERT

1 Allows DataStage plug-in stages to handle Oracle databases which use the special characters # and $ in column names.

6.
Environment Variable
$APT_TERA_SYNC_DATABASE $APT_TERA_SYNC_USER $APT_TER_SYNC_PASSWORD $APT_TERA_64K_BUFFERS

Teradata Environment Variables
Setting
[name] [user] [password] 1 1

Description
Starting with v7, specifies the database used for the terasync table. By default, EE uses the Starting with v7, specifies the user that creates and writes to the terasync table. Specifies the password for the user identified by $APT_TERA_SYNC_USER. Enables 64K buffer transfers (32K is the default). May improve performance depending on network configuration. This environment variable is not recommended for general use. When set, this environment variable may assist in job debugging by preventing the removal of error tables and partially written target table. Disables permission checking on Teradata system tables that must be readable during the TeraData Enterprise load process. This can be used to improve the startup time of the load.

$APT_TERA_NO_ERR_CLEANUP

$APT_TERA_NO_PERM_CHECKS

1

7.
Environment Variable
$APT_MONITOR_TIME

Job Monitoring Environment Variables
Setting [seconds] Description In v7 and later, specifies the time interval (in seconds) for generating job monitor information at runtime. To enable size-based job monitoring, unset this environment variable, and set $APT_MONITOR_SIZE below. Determines the minimum number of records the job monitor reports. The default of 5000 records is usually too small. To minimize the number of messages during large job runs, set this to a higher value (for example, 1000000). Disables job monitoring completely. In rare instances, this may improve performance. In general, this should only be set on a per-job basis when attempting to resolve performance bottlenecks. Prints record counts in the job log as each operator completes processing. The count is per operator per partition.

$APT_MONITOR_SIZE

[rows]

$APT_NO_JOBMON

1

$APT_RECORD_COUNTS

1

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

157 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

8. PerformanceTuning Environment Variables
Environment Variable
$APT_BUFFER_MAXIMUM_MEMORY

Setting 41903040 (example)

$APT_BUFFER_FREE_RUN

1000 (example)

Description Specifies the maximum amount of virtual memory, in bytes, used per buffer per partition. If not set, the default is 3MB (3145728). Setting this value higher will use more memory, depending on the job flow, but may improve performance. Specifies how much of the available in-memory buffer to consume before the buffer offers resistance to any new data being written to it. If not set, the default is 0.5 (50% of $APT_BUFFER_MAXIMUM_MEMORY). If this value is greater than 1, the buffer operator will read $APT_BUFFER_FREE_RUN * $APT_BUFFER_MAXIMIMUM_MEMORY before offering resistance to new data. When this setting is greater than 1, buffer operators will spool data to disk (by default scratch disk) after the $APT_BUFFER_MAXIMUM_MEMORY threshold. The maximum disk required will be
$APT_BUFFER_FREE_RUN * # of buffers * $APT_BUFFER_MAXIMUM_MEMORY

$APT_PERFORMANCE_DATA

directory [path]

$TMPDIR

Enables capture of detailed, per-process performance data in an XML file in the specified directory. Unset this environment variable to disable. Defaults to /tmp. Used for miscellaneous internal temporary data including FIFO queues and Transformer temporary storage. As a minor optimization, may be best set to a filesystem outside of the DataStage install directory.

9.
Environment Variable
$OSH_PRINT_SCHEMAS

Job Flow Debugging Environment Variables
Setting
1

Description
Outputs the actual schema definitions used by the DataStage EE framework at runtime in the DataStage log. This can be useful when determining if the actual runtime schema matches the expected job design table definitions. Disables operator combination for all stages in a job, forcing each EE operator into a separate process. While not normally needed in a job flow, this setting may help when debugging a job flow or investigating performance by isolating individual operators to separate processes. Note that disabling operator combination will generate more UNIX processes, and hence require more system resources (and memory). Disabling operator July 17, 2006 158 of 179

$APT_DISABLE_COMBINATION

1

The Advanced Stage Properties editor in DataStage Designer v7.1 and later allows combination to be enabled and disabled for on a per-stage basis.

Parallel Framework Red Book: Data Flow Job Design

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence

$APT_PM_PLAYER_TIMING

1 1 FORCE

$APT_PM_PLAYER_MEMORY $APT_BUFFERING_POLICY

combination also disables internal optimizations for job efficiency and run-times. Prints detailed information in the job log for each operator, including CPU utilization and elapsed processing time. Prints detailed information in the job log for each operator when allocating additional heap memory. Forces an internal buffer operator to be placed between every operator. Normally, the DataStage EE framework inserts buffer operators into a job flow at runtime to avoid deadlocks and improve performance. Using $APT_BUFFERING_POLICY=FORCE in combination with $APT_BUFFER_FREE_RUN effectively isolates each operator from slowing upstream production. Using the job monitor performance statistics, this can identify which part of a job flow is impacting overall performance. Set this environment variable to capture copies of the job score, generated osh, and internal Enterprise Edition log messages in a directory corresponding to the job name. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. This environment variable should not normally need to be set. When trying to start very large jobs on heavilyloaded servers, lowering this number will limit the number of processes that are simultaneously created when a job is started. For heavily loaded MPP or clustered environments, this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. The default is 30 seconds.

Setting
$APT_BUFFERING_POLICY=FORCE is not

recommended for production job runs.
$DS_PX_DEBUG

1

$APT_PM_STARTUP_CONCURRENCY

5

$APT_PM_NODE_TIMEOUT

[seconds]

Parallel Framework Red Book: Data Flow Job Design

July 17, 2006

159 of 179

© 2006 IBM Information Integration Solutions. All rights reserved. No part of this publication may be reproduced, transmitted, transcribed, stored in a retrieval system, or translated into any language in any form by any means without the written permission of IBM.

Information Integration Solutions Center of Excellence Appendix F: Sorting and Hashing Advanced Example The standard recipe for using the ‘Inter-Record Relationship Suite’ (Sort. record adjacency. c) Sort clusters and orders Sorting is often overkill. In most cases. a. Sorting is rarely required by the business logic. transcribed. at the same time. key-clustering is sufficient in many cases (a-only. below). will notice that two rows have identical values in the user-defined key column only if the two rows are contiguous. Sort actually does two things: (i) Groups rows that share the same values in key columns (forces related rows to be contiguous. stored in a retrieval system. b) Hash gathers Hash gathers into the same partition. . Merge. but there is little you Parallel Framework Red Book: Data Flow Job Design July 17. The reason for this requirement lies in the “light-weight” nature of these stages. it does not care about the key value of the next cluster with respect to the current key value—in part because this stage takes only one input. Join. 2006 160 of 179 © 2006 IBM Information Integration Solutions. a) Sort within partitions. transmitted. all require sorted inputs. across all partitions. sorting is needed to satisfy an input requirement of a downstream stage. Merge. but is frequently inefficient as records are ‘over-hashed’ and ‘over-partitioned’.k.a. in the same order.a. All rights reserved. this may be review for you. This approach is guaranteed to work. and related stages) is: Hash and Sort/Join/Merge on exactly the same keys. for example.k. The second portion of this Appendix assumes you have read and thoroughly understand these concepts. partition-wise co-location).) on any super-set of the keys.e. not globally. only needs to see two records at a time —one from each input stream—to do its job. etc. It will be followed by a detailed example that discusses these ideas in much greater depth. Join. No part of this publication may be reproduced. An illustrative piece of information. rows from all partitions that share the same value in key columns. for example. for example. related rows are in the same partition. i. RemoveDuplicates. key clustering. Remove Duplicates requires only (i): when it completes processing all the rows in a key cluster. These operations take place in parallel. There is also an “advanced” rule: a) b) Hash on any sub-set of the keys Sort (join/etc. but other rows may separate them within that partition. or translated into any language in any form by any means without the written permission of IBM.) (ii) Orders the clusters resulting from (i). in order This Appendix contains descriptions of what happens “behind the scenes”. If you have a lot of experience with hashing and sorting. Join. This creates partition-wise concurrency (a. RemoveDuplicates.

in part. work in parallel. Partitioners. Whenever you re-partition your sorted data. No part of this publication may be reproduced. and your data will retain its previous sort order. follow the partitioner with a PSM. or translated into any language in any form by any means without the written permission of IBM. . one from each input. Join and Merge. to the fact these stages take multiple input links—these stages only see 2 records at a timei. Inside a partitioner In Enterprise Edition. clustering. This due. there are more advanced methods to sort and partition that can leverage this capability and mitigate the cost of sorting vs. If the values on both inputs aren't ordered (vs. this is one instance where this might be handy). as a rule. they DO care about the key value of the next cluster. require both (i) and (ii). All rights reserved. 2006 161 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence can do to take advantage of it as there are no stages which guarantee key-clustering but do not perform a sort (some databases might be able to do key-clustering more cheaply than a sort. In other words. There is a component that will allow you to partition sorted data and achieve a sorted result: parallelsortmerge (PSM). any existing sort order is usually destroyed—see example below. When this stage completes processing all the rows in a key cluster. (and most other stages) do not gratuitously alter existing intra-partition row order. On multi-partitioned (i. Enterprise Edition will not. on the other hand. To restore row-adjacency.e. a sort operation is needed even on previously sorted columns following any partitioner. Enterprise Edition itself normally manages this use of this component. allow a row in a partition to jump ahead of another row in the same partitionii. Note that neither partition has a sorted result despite P1 having a sorted input (read row order from the bottom up. ‘streaming-style’). reshuffle rows among partitions. See usage notes in footnotesiii. Join/Merge can't effectively choose which input stream to advance for finding subsequent key matches.. but it can be invoked via the generic stage. partitioners (except for SAME). non-sequential) inputs. transmitted. grouped/clustered for remove duplicates). so row order between the two inputs is obviously critical. partitioners. Partitioner P0 2 101 P1 1 103 July 17. like stages. stored in a retrieval system. d) Partitioners respect row order but split clusters. for example: Repartitioning: P0 2 1 3 P1 103 102 101 Note that ‘1’ and ‘101’ have switched partitions. however. As you will see in this Appendix. Nonetheless. transcribed. 1.

Here is another possible outcome: Also: Consider the result of running the same job with the same data. transmitted. but a different number of partitions. No part of this publication may be reproduced. Minimizing Record Movement for Maximizing Performance Now we have covered the basic rules and mechanics for hash-partitioning and sorting. 6 There is an exception to this rule: If your hash key has only one value. Illustrated Above Before Partitioner Partition 0 3 Eve Smith Pine 3 Eve 1 Orlando Jones Elm 2 Adam 2 Adam Smith Pine 10 Rose 1 Partition 1 10 Rose Jones Pine 1 Orlando 1 10 Boris Smith Walnut 10 Boris 2 2 10 John Zorn Walnut 10 John 3 3 After Hash partitioning on Street/Tree: Illustrated Above After Partitioner Partition 0 Smith 3 Eve Smith Pine 3 Eve Smith 10 Rose Jones Pine 1 Orlando 1 Jones 2 Adam Smith Pine 10 Rose 1 Partition 1 Jones 10 Boris Smith Walnut 10 Boris 2 2 Smith 10 John Zorn Walnut 10 John 3 Zorn3 1 Orlando Jones Elm 2 Adam Smith Jones Jones Smith Zorn Smith There is more than one way to correctly hash-partition any Data Set6. 2006 162 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. transcribed. July 17.Information Integration Solutions Center of Excellence 3 102 Example: 6 rows in 2 partitions. Partition 0 Partition 1 1 Orlando 10 Rose 1 10 Boris 2 2 Adam 10 3 3 John Eve Jones Jones Elm Pine 2 1 Adam Smith Orlando Jones Boris Rose John Eve Smith Jones Zorn Smith Smith Walnut 10 2 Smith Pine 10 1 Zorn Walnut 10 3 Smith Pine 3 2. All rights reserved. . or translated into any language in any form by any means without the written permission of IBM. let’s look at how we can capitalize on these behaviors for performance benefits. stored in a retrieval system.

This would make it possible to determine how well each store is doing in relation to the national averages and track these performance trends over time. and append these values to the original data. 2006 163 of 179 © 2006 IBM Information Integration Solutions. They wish to determine the weighted average transaction amount peritem nation-wide. No part of this publication may be reproduced. There are many common extensions on gathering these kinds of sales metrics that take the following ideas and increase the scale of the problem at hand. per store for all stores in the nation. stored in a retrieval system. . Here is our source data: Data Set 1: 32 Rows Store Location 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Item ID 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 Transaction Date 2004-01-01 2004-01-02 2004-01-01 2004-01-03 2004-01-04 2004-01-02 2004-01-04 2004-01-03 2004-01-04 2004-01-03 2004-01-01 2004-01-04 2004-01-02 2004-01-02 2004-01-01 2004-01-03 2004-01-03 2004-01-01 2004-01-02 2004-01-04 2004-01-02 2004-01-03 2004-01-01 2004-01-04 2004-01-04 2004-01-01 2004-01-03 2004-01-02 2004-01-04 2004-01-02 2004-01-01 2004-01-03 Transaction Amt 1 2 3 5 5 54 7 8 2 3 45 65 7 85 9 98 23 3 32 45 54 56 7 8 23 45 534 6 65 7 78 87 The screen capture below shows how to implement the business logic in an efficient manner. thereby increasing the value of this exercise.Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets nation-wide. transcribed. taking advantage of Enterprise Edition’s ability to analyze a jobflow and insert sorts and partitioners in Parallel Framework Red Book: Data Flow Job Design July 17. or translated into any language in any form by any means without the written permission of IBM. transmitted. as well as the average transaction amount per-item. All rights reserved.

Here. This is the nation-wide transaction average per item (weighted by transaction. To enable automatic sort insertion. The Aggregator StoreAverageItemTransactionAmt will aggregate the data on ‘Store ID’. or translated into any language in any form by any means without the written permission of IBM. you will disable this facility. To do this. stored in a retrieval system. ensure that the environment variable. transmitted. so you want to leave them enabled (the default). with ANY value. and ‘Transaction Date’. with ANY value. . you will disable this facility. Enterprise Edition will hash-partition and sort on ‘Item ID’. ensure that the environment variable. and calculate the average of the ‘Transaction Amt’ column and place the results in a column named ‘Store Average Item Transaction Amt’. to get the original data with the averages appended. transcribed. This is the per-store transaction average per item. automatic sort insertion. 2006 164 of 179 © 2006 IBM Information Integration Solutions. and automatic partition insertion must be turned on7. ‘Item ID’. we will need to join each aggregator’s output back to the original data in order to get the original row count. All rights reserved. Here you want to let DS/EE choose where to insert sorts and partitioners for you. and ‘Transaction Date’. Enterprise Edition will hash-partition and sort on ‘Store ID’. Parallel Framework Red Book: Data Flow Job Design July 17. both on the output link of JoinSourceToAggregator_1): NOTE: In this job. Since the aggregator reduces row count (to the group count). The Aggregator stage NationalAverageItemTransactionAmt will aggregate the data on ‘Item ID’. not store). APT_NO_PART_INSERTION is NOT defined in your environment. and ‘Transaction Date’ calculating the average of the ‘TransactionAmt’ column and place the results in a column named ‘National Average Item Transaction Amt’. No part of this publication may be reproduced. there is only one sort and one repartition in the diagram. and ‘Transaction Date’. If you allow this environment variable to exist with any value. This is done with JoinSourceToAggregator_1 and JoinSourceToAggregator_2. If you allow this environment variable to exist with any value. APT_NO_SORT_INSERTION is NOT defined in your environment. ‘Item ID’. 7 To enable automatic partition insertion.Information Integration Solutions Center of Excellence appropriate places notice.

transcribed. or translated into any language in any form by any means without the written permission of IBM. The sort method requires the input data to be hashed and sorted. It does this by keeping running totals in memory for the aggregation for each output group.5 23 23 23 23 26.5 35. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’. The hash method only requires the input data to be hashed.25 16.25 26.5 35.5 180.25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17. however. not Hash.5 310. Therefore.5 6. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. it guarantees the output order to be sorted since the result of aggregation can be released for downstream processing as soon as the key change is detected.25 26. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation.5 3.5 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 56 23 85 7 6 7 3 98 87 534 PeekFinalOutput.5 180.5 3.5 50. it does not guarantee output order.5 Store Average Item Transaction Amt 28 28 43 43 6.5 3. stored in a retrieval system. No part of this publication may be reproduced.5 35. .5 310.5 46 46 6. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 180. All rights reserved.5 39. in return.5 6.Information Integration Solutions Center of Excellence NOTE: For this scenario. transmitted. you will need to set the Aggregator’s “Method” to Sort.25 26.5 16. 2006 165 of 179 1 1 2 2 1 1 1 1 1 1 1 1 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 1 3 3 7 7 5 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions.5 50.25 180.5 39. and ‘Average Item Transaction Amt By Store’) The output Data Set should look something like this (A 3-node configuration file was used in this implementation): Data Set 2: 32 Rows PeekFinalOutput.

transcribed. Parallel Framework Red Book: Data Flow Job Design July 17.75 38. more efficientv.25 44.e.75 38. .5 33. 6.Information Integration Solutions Center of Excellence 2 2 1 1 2 2 1 1 2 2 1 1 2 2 2 2 2 2 2 2 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 8 45 45 9 78 45 2 65 65 23 16.25 16. No part of this publication may be reproduced. This is certainly an improvement on the previous answer. Here’s a screen shot of this.75 26. as only partition concurrency is affected by sequential execution.25 44.25 44. even sequential job execution does not alter the answer for the sort-related questions. 3.5 33.5 44 44 Since both the Aggregator and Join expect the data to arrive hashed and sorted on the grouping key(s) —both operations that consume large amounts of CPU—a couple of questions arise with respect to efficiency: What is the minimum number of hash-partitioners needed to implement this solution correctly? What is the minimum number of sorts needed to implement this solution? What is the minimum number of times that sort will need to buffer the entire Data Set to disk to implement this solution? Though running the job sequentially eliminates questions related to partitioners.25 38. A deeper examination (of the score dump. An examination of the job above would suggest: 6. transmitted.: record adjacency assumes partition concurrency8. 3. i.75 38. or translated into any language in any form by any means without the written permission of IBM. and 6. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition. 2006 166 of 179 © 2006 IBM Information Integration Solutions.25 44.5 27 27 61. All rights reserved. stored in a retrieval system. and 3. and 1.5 61. appended to the end of this document for masochistsiv) might suggest: 4. A much better answer is: 1.

The settings in the sort should look like this: Parallel Framework Red Book: Data Flow Job Design July 17. we still need to fix the sort order. but we know that the data is already sorted on TransactionDate. we hash and sort on ItemID and TransactionDate only. and ItemId. This combination of hash and sort adequately prepares the data for NationalAverageItemTransactionAmt.Information Integration Solutions Center of Excellence NOTE: In this job. automatic sort insertion. However. in order. Hashing on these fields will gather all unique combinations of ItemID and TransactionDate into the same partition. What is wrong with the data? The sort order does not include the StoreLocation. You may be wondering why the partitioning wasn’t mentioned as part of the problem. 2006 167 of 179 © 2006 IBM Information Integration Solutions. One would expect that we would need to sort on StoreLocation. No part of this publication may be reproduced. as it isn’t available on the link sort. . In our initial copy stage (DistributeCopiesOfSourceData). and automatic partition insertion must be turned offvii. transcribed. or translated into any language in any form by any means without the written permission of IBM. and ItemId. but you must use the sort stage to access it. TransactionDate. The ‘advanced’ rule for hash partitioning is: you may partition on any sub-set of the aggregation/join/sort/etc. Sort offers an efficiency-mode for pre-sorted data. However. just as in the previous example. This is a problem for StoreAverageItemTransactionAmt. the data is not properly prepared for StoreAverageItemTransactionAmt. This is because the data is already partitioned in a compatible manner for the aggregator. stored in a retrieval system. but it is a lengthy parenthetical statement that would interrupt the flow of the scenario discussion). as it expects all of the records for a particular StoreLocation/TransactionDate/ItemId combination to arrive on the same partition. transmitted. keys (viii This footnote contains key concepts that this document addresses. All rights reserved.

Getting back to the aggregators. This means that the output of the DistributeCopiesOfSourceData and NationalAverageItemTransactionAmt are already hashed and sorted on the keys needed to perform JoinSourceToAggregator_1. but the group keys will force the proper order). especially in MPP environments where repartitioning implies network I/O). even for ‘fast’ disks). ALL extant values of the remaining columns are already in the same partition. This accomplishes the first goal. However since we sub-sorted the data before this aggregator (unlike NationalAverageItemTransactionAmt). 2006 168 of 179 © 2006 IBM Information Integration Solutions.Information Integration Solutions Center of Excellence As you can see. If the group size was only several hundred records. to append a column representing the national (weighted) average item transaction amount. instead of buffering the entire Data Set. or translated into any language in any form by any means without the written permission of IBM. Since the aggregator does not need to disturb row-order (for pre-sorted data). however dreadful). . The previous two paragraphs contain two key concepts in Enterprise Edition (pun fully intended. we will have to prep the output from the first join to account for Parallel Framework Red Book: Data Flow Job Design July 17. we have instructed the sort stage that the data is already sorted on ItemID and TransactionDate (as always with sorting records. The output of StoreAverageItemTransactionAmt contains the other column we need to append to our source rows. This lets sort know that it only needs to gather all records with a unique combination of ItemID and TransactionDate in order to sort a batch of records. granted. but the entire Data Set was 100 million records. transmitted. stored in a retrieval system. the rows will come out in the same order they went in (different rows. All rights reserved. Also worth noting here: because we already hashed the data on ItemID and TransactionDate. key order is very important). No part of this publication may be reproduced. this would save a tremendous amount of very expensive disk I/O as sort can hold a few hundred records in memory in most cases (disk I/O is typically several orders of magnitude more costly than memory I/O. and we only want to ‘sub-sort’ the data on StoreLocation (this option is only viable for situations where you need to maintain the sort order on the initial keys). which is what makes this sort possible w/o re-partitioning (which is also quite expensive. transcribed.

5 3. No part of this publication may be reproduced.5 16.25 44. or translated into any language in any form by any means without the written permission of IBM. as it is much more expensive than non-stable sorts.5 33.75 38. Output from above solution: Data Set 3: PeekFinalOutput.5 3.25 44.5 26. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 3. transcribed. and it is enabled by default.25 44.5 16.Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt.75 38. All rights reserved. stored in a retrieval system. useful for preserving previous sort orderings).75 Store Average Item Transaction Amt 2 2 6 6 5 5 26. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17.25 16.5 61. .5 33.5 44 44 1 1 1 1 2 2 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 3 1 5 7 3 7 45 8 9 45 45 78 2 65 23 65 PeekFinalOutput.25 44.75 38.25 38.5 27 27 61. 2006 169 of 179 © 2006 IBM Information Integration Solutions.25 16. transmitted. This sort will look exactly like the other sort stage: Remember to disable the ‘Stable Sort’ option if you do not need it (it will try to maintain row order except as needed to perform the sort.

a savings of 400.5 46 46 6.5 50.5 180. With the initial solution. a single partitioner.5 This solution produces the same result but is achieved with only one complete sort.5 35. .25 26.25 26. and only hashes.Information Integration Solutions Center of Excellence Store Location Item ID Transaction Date Transaction Amt 1 1 2 2 1 1 1 1 2 2 1 1 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1 1 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 7 85 7 6 3 98 23 56 534 87 National Average Item Transaction Amt 35.5 23 23 26. transmitted.25 26. 100.5 39. but there is a critical difference.25 180.000.5 35. It looks very similar to the first solution.000. transcribed. or translated into any language in any form by any means without the written permission of IBM.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume. Imagine a job with 100 million records as the input.5 39. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17.5 35. All rights reserved.5 6. stored in a retrieval system.5 Store Average Item Transaction 28 28 43 43 6.000. That is a LOT of saved processing power.000 records.5 180. we had to sort (on disk) 300.5 50.5 310.000 records. 2006 170 of 179 © 2006 IBM Information Integration Solutions.5 6.000 records in addition to hashing 300. 000. and two sub-sorts—a much more efficient solution for large data volumes.5 23 23 180.000.000 records.5 310. There is an even more efficient solution. The second solution only sorts (on disk) 100.

25 16.Information Integration Solutions Center of Excellence NOTE: In this job. Data Set 4: PeekFinalOutput.25 16. However it will not create additional processes to handle the records and re-order them.5 16. Looks a lot like solution 1. transmitted. we have chosen to use the StoreLocation column as a part of our sorting key. except w/o the sort on the output of JoinSourceToAggregator.5 3.5 3. The difference is on DistributeCopiesOfSourceData: Here. This is a potentially huge savings on large data volumes (remember the previous example).5 3. automatic sort insertion. Also. transcribed. 2006 171 of 179 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 1 1 2 2 2004/01/01 2004/01/01 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 2004/01/01 2004/01/01 1 3 7 3 5 7 8 45 45 9 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. for the same reasons. Comparing the efficiency of this solution with that of number two.25 44. we saved a sub-sort on 100 million records . Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 16. stored in a retrieval system.25 44. All rights reserved. Here is the output from this version of the job.a significant savings. This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26. .5 26. No part of this publication may be reproduced. or translated into any language in any form by any means without the written permission of IBM. the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed. but NOT to use it for hashing.5 27 27 July 17. and automatic partition insertion must be turned off vii.

5 33.5 180.5 23 23 23 23 26. inhibits pipe-lining (by buffering large amounts of data to disk since it needs to see all data before it can determine the resulting sorted sequence)ix.75 38.25 26.75 61.25 180.5 310.25 38.25 26.5 310. 2006 172 of 179 © 2006 IBM Information Integration Solutions.25 44.75 38. transcribed. stored in a retrieval system.75 38. transmitted. in addition to the heavy penalty paid in disk I/O for using a full sort.5 50.5 Store Average Item Transaction Amt 28 28 43 43 6.5 180. Here is a screen shot of a sort running on 40 million records: Parallel Framework Red Book: Data Flow Job Design July 17.5 6.5 39. No part of this publication may be reproduced.5 46 46 6. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 6. by definition.5 44 44 PeekFinalOutput. sort.5 35. or translated into any language in any form by any means without the written permission of IBM.5 35.5 39.5 33.5 35.5 1 1 2 2 1 1 2 2 1 1 2 2 1 1 2 2 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2004/01/02 2004/01/02 2004/01/02 2004/01/02 2004/01/03 2004/01/03 2004/01/03 2004/01/03 2 54 54 32 5 8 23 56 7 85 7 6 3 98 534 87 Finally.5 180.25 26. .Information Integration Solutions Center of Excellence 2 2 1 1 2 2 2 2 2 2 2 2 2004/01/01 2004/01/01 2004/01/04 2004/01/04 2004/01/04 2004/01/04 78 45 2 65 23 65 44.5 50. All rights reserved.

although ~5 million records have entered the sort. instead of waiting until all 40 million records have been sorted (in this instance. requiring a large amount of scratch diskx. transmitted. All rights reserved.Information Integration Solutions Center of Excellence As you can see. transcribed. we are sub-sorting the data we sorted in the previous diagram). This situation is analogous to all of the sorts in solution 1. This allows down-stream stages to be processing data during the sorting process. you can clearly see that a sub-sort does not inhibit pipe-lining--very nearly the same number of rows have entered and left the sort stage (and NO buffering is required to perform the subsort). stored in a retrieval system. No part of this publication may be reproduced. . and the link sort in solution three. 2006 173 of 179 © 2006 IBM Information Integration Solutions. Here is an example of a ‘sub-sort’: Here. or translated into any language in any form by any means without the written permission of IBM. This is because a standard sort requires all rows to be present in order to release the first row. the link sort in solution 2. no rows have left yet. Parallel Framework Red Book: Data Flow Job Design July 17.

0 or larger. At least one of these dataset must be re-hashed. without curly braces (mutually exclusive: -ci.property=value pair(s). key={ value=TransactionDate } })#>eCollectAny op2[3p] (parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt)} ds2: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }. problem: you created a fileset with 8 nodes.descending sort order. Join needs to see all of the rows in the current cluster on at least one of the input links.case-insensitive comparison.case-sensitive comparison. but the job that reads it only has 4 nodes. the sort order will be preserved. value one of first.dataset to be resorted/merged. optional -cs -. optional This operator may have following inputs -Sorted -.queue length at which to issue a warning. optional -param -. last. You could hash only on A. exactly one occurrence required This operator may have following outputs -reSorted -. but not introduce superfluous sorts. presorted.input field name Sub-options for key: -ci -. In other cases. default -nulls -. default -desc or descending -. exactly one occurrence required <add example here of how psm works> iv Dump score for solution 1 main_program: This step has 16 datasets: ds0: {op0[3p] (parallel SourceData. iii ParallelSortMerge Operator Options: -key -. but suppose that A has too small a number of unique values (country codes. optional params -. otherwise a Cartesian product is impossible. which would result in having to completely re-sort that dataset despite having a sorted version already. default=10000 -doStats -. datasets with differing partition counts and you wish to join/merge them. optional.DSLink2) eAny=>eCollectAny op1[3p] (parallel DistributeCopiesOfSourceDta)} ds1: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }. key={ value=TransactionDate } })#>eCollectAny op4[3p] (parallel APT_TSortOperator(0))} ds4: {op2[3p] (parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) [pp] eSame=>eCollectAny op5[3p] (parallel APT_TSortOperator(1))} ds5: {op3[3p] (parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) [pp] eSame=>eCollectAny op8[3p] (parallel APT_TSortOperator(2))} ds6: {op4[3p] (parallel APT_TSortOperator(0)) . race/gender/ethnicity codes are typical).where null values should sort. There are other situations where this is valuable but they are much less common. This ‘problem’ is addressed by the parallelsortmerge component iii. Another common problem: You need to Hash and Sort on columns A and C to implement the business logic in one section of the flow. key={ value=StoreLocation } })#>eCollectAny op3[3p] (parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt)} ds3: {op1[3p] (parallel DistributeCopiesOfSourceDta) eOther(APT_HashPartitioner { key={ value=ItemID }. 1 or more name -.extra parameters for key. optional position -. -cs) (mutually exclusive: -asc.specify a key field. it’s only true for cases where the key is unique.ascending sort order. 2147483647 or smaller. -desc) -warnLevel -. Normally EE would re-partition the data into 4 nodes + destroy your sort order. However. default=first -asc or ascending -. A third. gender codes. less common. you can use the ParallelSortMerge stage to ensure that no matter the degree of parallelism of the writer + reader.string. optional. optional records -.integer. This would allow you to combine other columns into your hash key to reduce data-skew.i This is an over-simplification. ii A common problem: Suppose you have two (or more).resorted dataset. key={ value=TransactionDate }.report statistics at the end of the run. but in another section you need to hash and sort on columns A and B.

p1] node3[op3. subArgs={ cs } } })#>eCollectAny op10[3p] (parallel JoinSourceToAggregator_2.p1] node3[op5.p0] node1[op0.p2] )} op5[3p] {(parallel APT_TSortOperator(1)) on nodes ( node1[op5.p0] node2[op3.p0] node1[op1.p1] node1[op1.p0] node2[op2.p1] node1[op0.DSLink18_Sort)} ds12: {op10[3p] (parallel JoinSourceToAggregator_2.DSLink18_Sort) [pp] eSame=>eCollectAny op11[3p] (parallel buffer(2))} ds13: {op11[3p] (parallel buffer(2)) [pp] eSame=>eCollectAny op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds14: {op12[3p] (parallel buffer(3)) [pp] eSame=>eCollectAny op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds15: {op13[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2) [pp] eSame=>eCollectAny op14[3p] (parallel PeekFinalOutput)} It has 15 operators: op0[3p] {(parallel SourceData.DSLink2) on nodes ( node1[op0. subArgs={ cs } }.[pp] eSame=>eCollectAny op6[3p] (parallel buffer(0))} ds7: {op5[3p] (parallel APT_TSortOperator(1)) [pp] eSame=>eCollectAny op7[3p] (parallel buffer(1))} ds8: {op6[3p] (parallel buffer(0)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds9: {op7[3p] (parallel buffer(1)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds10: {op8[3p] (parallel APT_TSortOperator(2)) [pp] eSame=>eCollectAny op12[3p] (parallel buffer(3))} ds11: {op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1) [pp] eOther(APT_HashPartitioner { key={ value=ItemID.p2] )} op6[3p] {(parallel buffer(0)) .p2] )} op4[3p] {(parallel APT_TSortOperator(0)) on nodes ( node1[op4.p2] )} op2[3p] {(parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op2.p1] node3[op4.p2] )} op3[3p] {(parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op3. key={ value=StoreLocation. key={ value=TransactionDate }.p0] node2[op5.p0] node2[op4.p1] node3[op2.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op1.

p2] )} op9[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_1) on nodes ( node1[op9.e.p2] )} op7[3p] {(parallel buffer(1)) on nodes ( node1[op7.p1] node3[op7.p2] )} It runs 45 processes on 3 nodes. if you move records unnecessarily.p1] node3[op12.p2] )} op10[3p] {(parallel JoinSourceToAggregator_2. key={ value=TransactionDate } })#>eCollectAny op1[3p] (parallel DistributeCopiesOfSourceDta.p1] node3[op6.p0] node2[op9.p0] node2[op10.p2] )} op8[3p] {(parallel APT_TSortOperator(2)) on nodes ( node1[op8.p2] )} op13[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_2) on nodes ( node1[op13. or order.p0] node2[op11.p0] node2[op12.p1] node3[op13. v Throughout this document the general meaning of the phrase ‘more efficient’ is fewer record movements--i.p0] node2[op6.p0] node2[op7.p2] )} op12[3p] {(parallel buffer(3)) on nodes ( node1[op12.p1] node3[op11.p0] node2[op14.p2] )} op14[3p] {(parallel PeekFinalOutput) on nodes ( node1[op14. a record changes partition.p1] node3[op9. fewer times.p2] )} op11[3p] {(parallel buffer(2)) on nodes ( node1[op11.p0] node2[op13.on nodes ( node1[op6. vi Dump Score for Solution 2 main_program: This step has 15 datasets: ds0: {op0[3p] (parallel SourceData. subArgs={ cs } }.DSLink2_Sort) [pp] eSame=>eCollectAny op2[3p] (parallel DistributeCopiesOfSourceDta)} ds2: {op2[3p] (parallel DistributeCopiesOfSourceDta) .DSLink2_Sort)} ds1: {op1[3p] (parallel DistributeCopiesOfSourceDta.p1] node3[op14.p1] node3[op10.p0] node2[op8.DSLink18_Sort) on nodes ( node1[op10.DSLink2) eOther(APT_HashPartitioner { key={ value=ItemID.p1] node3[op8. Since moving records around takes CPU time and extra system calls. your run time will be adversely affected.

p0] node2[op2.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta.DSLink2) on nodes ( node1[op0.p0] node2[op5.DSLink2_Sort) on nodes ( node1[op1.[pp] eSame=>eCollectAny op3[3p] (parallel SubSortOnStoreLocation)} ds3: {op2[3p] (parallel DistributeCopiesOfSourceDta) [pp] eSame=>eCollectAny op4[3p] (parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt)} ds4: {op2[3p] (parallel DistributeCopiesOfSourceDta) [pp] eSame=>eCollectAny op6[3p] (parallel buffer(0))} ds5: {op3[3p] (parallel SubSortOnStoreLocation) [pp] eSame=>eCollectAny op5[3p] (parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt)} ds6: {op4[3p] (parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) [pp] eSame=>eCollectAny op7[3p] (parallel buffer(1))} ds7: {op5[3p] (parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) [pp] eSame=>eCollectAny op8[3p] (parallel buffer(2))} ds8: {op6[3p] (parallel buffer(0)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds9: {op7[3p] (parallel buffer(1)) [pp] eSame=>eCollectAny op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1)} ds10: {op8[3p] (parallel buffer(2)) [pp] eSame=>eCollectAny op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds11: {op9[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_1) [pp] eSame=>eCollectAny op10[3p] (parallel SubSortOnStoreLocation2)} ds12: {op10[3p] (parallel SubSortOnStoreLocation2) [pp] eSame=>eCollectAny op11[3p] (parallel buffer(3))} ds13: {op11[3p] (parallel buffer(3)) [pp] eSame=>eCollectAny op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2)} ds14: {op12[3p] (parallel APT_JoinSubOperator in JoinSourceToAggregator_2) [pp] eSame=>eCollectAny op13[3p] (parallel PeekFinalOutput)} It has 14 operators: op0[3p] {(parallel SourceData.p2] )} op3[3p] {(parallel SubSortOnStoreLocation) on nodes ( node1[op3.p1] node3[op2.p2] )} op5[3p] {(parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op5.p1] node1[op0.p0] node2[op3.p0] node2[op1.p1] node3[op4.p1] node3[op1.p2] )} op4[3p] {(parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op4.p2] .p0] node1[op0.p1] node3[op5.p1] node3[op3.p2] )} op2[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op2.p0] node2[op4.

p0] node2[op8.p2] )} op7[3p] {(parallel buffer(1)) on nodes ( node1[op7. and ColumnB: ColumnAColumnBColumnCGroup 1111112113Group 2121122123Group 3211212213Group 4221222223Group 5231232233 There must be exactly 5 groups identified by the hash algorithm b/c there are exactly 5 unique combinations of ColumnA and ColumnB. look at this example: Here is my source data: ColumnAColumnBColumnC111112113121122123211212213221222223231232233 Here are the possible outcomes if I hash-partition on ColumnA. Any combination of these groups can be in any partition. thus introducing the inefficiencies we are trying to avoid. vii In this instance.p2] )} It runs 42 processes on 3 nodes.p0] node2[op7.p1] node3[op7.p0] node2[op9.p0] node2[op10.p1] node3[op11. This does not mean that these groups will be in unique partitions: consider a job that only has 3 partitions. and the likelihood decreases with larger numbers of groups.p0] node2[op6.p0] node2[op12. in fact.p1] node3[op10. Here are the possible outcomes if I hash-partition on ColumnA only: ColumnAColumnBColumnCGroup 1111112113121122123Group 2211212213221222223231232233 .p2] )} op13[3p] {(parallel PeekFinalOutput) on nodes ( node1[op13.p0] node2[op13. the distribution of groups across partitions is nearly even for large numbers of groups).p1] node3[op6.p2] )} op9[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_1) on nodes ( node1[op9. you could have ALL 5 groups sent to the same partition (this is unlikely.p2] )} op11[3p] {(parallel buffer(3)) on nodes ( node1[op11. you want auto insertion turned off b/c EE will see that you are ‘missing’ a sort/partitioner and insert one for you.)} op6[3p] {(parallel buffer(0)) on nodes ( node1[op6.p1] node3[op8.p1] node3[op13.p0] node2[op11.p2] )} op12[3p] {(parallel APT_JoinSubOperator in JoinSourceToAggregator_2) on nodes ( node1[op12. viii To understand why this is true. regardless of the number of partitions: if you are running a job with 6 partitions.p2] )} op8[3p] {(parallel buffer(2)) on nodes ( node1[op8.p1] node3[op12.p1] node3[op9.p2] )} op10[3p] {(parallel SubSortOnStoreLocation2) on nodes ( node1[op10.

larger groups. we wouldn’t see any improvements in performance over a 2-way job. this grouping is OK. It is only true on a per-partition basis. . x This is a slight oversimplification. not for the entire dataset. even if we ran the job 12-ways. So hashing on fewer columns resulted in fewer. Therefore. You need to understand your data and make educated decisions about your hashing strategy. then all unique combinations of ColumnA and ColumnB are together. We can do this by identifying the intersection of keys needed among all of the hash-partitioners and hashing only on those keys: TransactionDate and ItemId NOTE: if you take this to an extreme. effectively. In the above example. and ColumnC. you will get a very small number of groups. One effect is that if we wanted to aggregate on ColumnA and ColumnB. we would have only two groups. consuming RAM and process space and offering nothing in return. we want to reduce the number of times that we hash (b/c partitioning costs CPU time). ColumnB. which will. as well as all unique combinations of ColumnA. summing ColumnC. ix This means that down-stream process will be sitting idle until the sort is completed. b/c: if all unique values of ColumnA are together. reduce the parallelism of the job. there are only two groups by hashing on ColumnA only. In the scenario that we are discussing in the main document.As you can see.

Sign up to vote on this title
UsefulNot useful