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.

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

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

No part of this publication may be reproduced. Performing transformation from data sources to data targets. databases and DS/EE Data Sets. stored in a Subroutine retrieval system. Performing row validation to support data quality. or translated into any language in any form by any means without the written permission of IBM.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. 2006 6 of 179 After Job © 2006 IBM Information Integration Solutions. 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. . and • Provisioning data targets. All rights reserved. transcribed. transmitted.

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

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

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

2006 10 of 179 © 2006 IBM Information Integration Solutions.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. . 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. transmitted. All rights reserved. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17.

No part of this publication may be reproduced. and the write-through cache records from the last load of the target database are merged. if the load fails. producing an ordered and linked associative table. The key column for a Referential Integrity check is validated by a Transformer stage. the cache is deleted. transcribed. In this case.2 Hybrid Jobs The following example hybrid job demonstrates several interesting techniques that might be used in more complex jobs. Parallel Framework Red Book: Data Flow Job Design July 17. stored in a retrieval system. 2006 11 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. By specifying a full-outer join we produce a Cartesian product dataset. we replicated the Oracle structure (lower input) for each country found in the write-through cache country dataset (upper input). transmitted. . Some of the more interesting solutions in this job are circled. 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). All rights reserved. The merged records are grouped and ordered before being de-duplicated to remove obsolete records.Information Integration Solutions Center of Excellence 1.2. 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. forcing other jobs that might depend on this data to access the existing (not updated) target database table. This job also loads the target database table and creates write-through cache. The non-validated records. The de-duplicated records are re-grouped and ordered before calculation of the terminating keys. the validated records. In this case. it is rejected by the transformer to a reject port and the validation is not performed for those records. If the key column is NULL.

stored in a retrieval system.3 Provisioning Jobs This example provisioning job demonstrates the straightforward approach to simple provisioning tasks.2. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. 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 1. . 2006 12 of 179 © 2006 IBM Information Integration Solutions. transmitted.

. / /DSEngine Data File Systems /PXEngine /Data0 /Configurations /Project_A . 2006 13 of 179 /Projects / /Project_A © 2006 IBM Information Integration Solutions.. These directories are configured during product installation. /DataN /P . Figure 1: RecommendedGigabyte 1 DataStage Install. stored in a retrieval system.1 Directory Structures 2. /Ascential /Scratch0 /ScratchN /Project_A /patches /P /DataStage /Project_Z .. transcribed..Information Integration Solutions Center of Excellence 2 Standards Establishing consistent development standards helps to improve developer productivity and reduce ongoing maintenance costs. Install. No part of this publication may be reproduced. and to build technical and support documentation... Development standards can also make it easier to integrate external processes such as automated auditing and reporting. 2.1... File systems are highlighted in blue /Project_Z . . Install FS Install File System Scratch File Systems . transmitted. and Data Directories /Project_Z Parallel Framework Red Book: Data Flow Job Design July 17.. All rights reserved.1 Data. Scratch.. 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.

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

transcribed. and production) as appropriate. or translated into any language in any form by any means without the written permission of IBM. /Project_Z Within the separate Staging file system. /Staging /dev Top-Level Directory development data tree. files are separated by Project name as shown below.. data directories are implemented for each deployment phase of a job (development. 2006 15 of 179 .. location of source data files. target data files.Information Integration Solutions Center of Excellence /Staging /dev /si /qa /prod /Project_A /Project_A /Project_A /Project_A /archive /archive /archive /archive .. No part of this publication may be reproduced. transmitted. Within each deployment directory.. 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 . /Project_Z July 17... /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. qa. ./archive of these development phases may be present on a local file not all system.. stored in a retrieval system. . error and reject files. If the file system is not /archive /archive /archive shared across multiple servers.. All rights reserved. system integration.

transmitted. which will enable DataStage applications to move through the life cycle without any code changes. July 17. such as the operating system. 2006 . stored in a No part of this /Project_Z /Project_Z retrieval system. through the use of environment variables. A directory structure should be created that organizes external entities and is directly associated with 1 and only 1 DataStage project. All rights reserved. The integration can be as simple as a file system for housing source files. 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. ..1./Project_Z publication may be reproduced.. transcribed. 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.. This will provide a convenient vehicle to group and manage resources used by a project. or it could require scripts for example integrating with an Enterprise Scheduler.... .Information Integration Solutions Center of Excellence 2.2 Extending the DataStage Project for External Entities It is quite common for a DataStage application to be integrated with external entities. /Project_Z . Environment variables are a critical portability tool. The directory structure will be made transparent to the DataStage application.. another application or middle ware.. 16 of 179 . 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. this directory may only be present in the /dev development code tree) location of DataSet header files (. No part of this publication may be reproduced. DataStage routines. 2006 17 of 179 © 2006 IBM Information Integration Solutions. qa. utilities. system integration. 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. 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. transmitted.Information Integration Solutions Center of Excellence Within the Project_Plus hierarchy.$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. If the file system is not shared across multiple servers. for example the following diagram shows Project_Plus variables as defined in the DataStage Administrator. or translated into any language in any form by any means without the written permission of IBM. not all of these development phases may be present on a local file system. transcribed. directories are created for each deployment phase of a job (development. a copy of dsenv and copies of DSParams. and production) as appropriate. stored in a retrieval system.ds file) location of custom job logs and reports location of parameter files for automated program control. . BuildOps. All rights reserved.

there may be external entities that are shared with other DataStage projects. for example in WebSphere MetaStage. stored in a retrieval system.Information Integration Solutions Center of Excellence Figure 4: Project_Plus Environment Variables In some implementations.2 Naming Conventions As a graphical development environment. the Designer tool assigns default names based on the object type. While the default names may create a functional data flow. 2. 2006 18 of 179 © 2006 IBM Information Integration Solutions. and the order the item is placed on the design canvas. 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. nor do they adequately document the business rules or subject areas. A similar directory structure to the Project_Plus structure could be configured and referred to as DataStage_Plus. By default. or an XML extract Parallel Framework Red Book: Data Flow Job Design July 17. 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. transcribed. they do not facilitate ease of maintenance over time. 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. All rights reserved. .

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

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

Shared Containers. dialog box fields that specify a new category have only one input area for defining the Category name. Category Names can be long. Routines. Therefore Directory names should be Word Capitalized and separated by either an underscore or a space. Table definitions…). are Alpha Numeric and can also contain both Spaces and Underscores. 2006 21 of 179 © 2006 IBM Information Integration Solutions. It may or may not have a one to one relationship with an organizations’ project of work. 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. allowing related objects to be grouped together.4 Naming Conventions by Object Type 2. This factor often can cause terminology issues especially in teamwork where both business and developers are involved.2.1 Projects Each DataStage Project is a standalone repository.2. Multiple levels of Hierarchy are named by specifying the Hierarchy levels separated by a backslash (“\”). No part of this publication may be reproduced. developers have the flexibility to define their own Directory or Category hierarchy.2 Category Hierarchy DataStage organizes objects in its repository by Categories. as well as documentation and auditing. which can provide impact analysis. However with the limit of 18 characters the name is most often composed of abbreviations.4. Test (“Test”).4.4. the structure “A Test\Lower\Lower Still” is shown below: Parallel Framework Red Book: Data Flow Job Design July 17.3 Job Category Naming Within Designer. All rights reserved. Below this level. and Production (“Prod”). 2. transmitted. Jobs. . stored in a retrieval system. The name of a DataStage Project may only be 18 characters in length. DataStage enforces the top level Directory Structure for different types of Objects (for example. or translated into any language in any form by any means without the written permission of IBM.2. 2. For example. it can contain alpha-numeric characters and it can contain underscores.Information Integration Solutions Center of Excellence processes such as WebSphere MetaStage.2. The suffix of a Project name should be used to identify Development (“Dev”). transcribed.

2006 22 of 179 © 2006 IBM Information Integration Solutions. Note that Job names must be unique within a DataStage project. No part of this publication may be reproduced. not within a category. For example. transcribed. a Category level should only contain objects that are directly related. 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. Where possible. with sub-levels for individual functional modules. stored in a retrieval system. Jobs and Job Sequences are grouped together in the same scope as the technical design documents. All rights reserved. a job category might contain a Job Sequence and all the jobs and only those jobs that are contained in that sequence. transmitted. or translated into any language in any form by any means without the written permission of IBM. all Jobs and Job Sequences will be grouped in a single parent Category. 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. For example.Information Integration Solutions Center of Excellence Figure 5: Creating Category Hierarchies The main reason for having Categories is to group related objects. Within each functional module category. . Categorization by Functional Module For a given application.

Remembering that Job names must be unique within a given project. transcribed. if these TableDefs are to be used by other jobs. On import. project manager. Although the default table definition categories are useful from a functional perspective. stored in a retrieval system. 2. DataStage assigns the level names based on the source of the metadata import (for example. care must be taken to override the default choices for category names during Table Definition import. they must be moved to the appropriate category and re-imported from that category in every job where they are used. All rights reserved. transmitted.4 Table Definition Categories Unlike other types of DataStage objects. categories will be created for each developer as their personal sandbox and place they perform unit test activities on jobs they are developing. The placement of these fields varies with the method of metadata import. and the responsibility of the development manager assigned the DataStage Manager role to ensure that projects are not obese with unused jobs.. It is the responsibility of each developer to delete unused or obsolete code. Orchestrate. PlugIn.2. 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. In the previous illustration. etc. Once created. Temporary TableDefs created by developers to assist with job creation appear under the Saved category by default. TableDefs that remain in the Saved category Parallel Framework Red Book: Data Flow Job Design July 17. When implementing a customized Table Definition categorization. By default. . 2006 23 of 179 © 2006 IBM Information Integration Solutions. Saved. or translated into any language in any form by any means without the written permission of IBM. 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. and there are 2 additional high-level categories. categories and metadata. but this can be overridden during import. establishing a Table Definition categorization that matches project development organization is recommended. No part of this publication may be reproduced. ECRP and Templates. Table Definitions are always categorized using two level names..4.). two developers have private categories for sandbox and development activities.Information Integration Solutions Center of Excellence Figure 6: Categorization by Functional Module Categorization by Developer In development projects. 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.

the TableDefs have been grouped into a master category of Custom. The following is one of the TableDefs from this project showing how to correctly specify the category and sub-category. This will be discussed further in Section 2. numbers. An alternative implementation is to set the “Data source name” to that of the source system or schema. Parallel Framework Red Book: Data Flow Job Design July 17. and underscores only.4. 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.g. transmitted. with sub-categories intended to identify the type of the source.5 Jobs and Job Sequences Job names must begin with a letter and can contain letters.4.Information Integration Solutions Center of Excellence should be deleted as soon as possible. transcribed. Figure 7: Table Definition Categories 2. e. No part of this publication may be reproduced.g: DWPH1 or ECRP. 2006 24 of 179 © 2006 IBM Information Integration Solutions. Because the name of can be long. stored in a retrieval system. Jobs and Job Sequences are all held under the Category Directory Structure of which the top level is the category “Jobs”.2 Category Hierarchy.2. . e. or translated into any language in any form by any means without the written permission of IBM.2. Each subject area will have a master category. In this example. All rights reserved.: Datasets. A Job will be suffixed with the class word “Job” and a Job Sequence will be suffixed with the class word “Seq”.

4.6 Shared Containers Shared containers have the same naming constraints as jobs in that the name can be long but can not contain underscores. It is particularly important to establish a consistent naming convention for link names. 2006 25 of 179 © 2006 IBM Information Integration Solutions. links are objects that represent the flow of data from one stage to the next.2. Examples of Parameter naming are: • Audit_Trail_Output_Path_Parm • Note where this is used in a stage property. or translated into any language in any form by any means without the written permission of IBM. When a Shared Container is used. To differentiate between Parallel Shared Containers and Server Shared Containers. transcribed.Information Integration Solutions Center of Excellence 2. Within the graphical Designer environment. Therefore the parameter name must be made readable using Capitalized words separated by underscores. during Link Ordering). stage editors identify links by name. It is optional as to whether you decide to change this code to something meaningful.8 Links Within a DataStage Job. links represent the flow of a message from one activity / step to the next.4.7 Parameters A Parameter can be a long name consisting of alphanumeric characters and underscores. the parameter name is delimited by the # sign: #Audit_Trail_Output_Path_Parm# 2. 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. a character code is automatically added to that instance of its use throughout the project. instead of using the default “DSLink#” (where “#” is an assigned number). All rights reserved.2.4. The class word suffix is “Parm”. so word capitalization should be used for readability. having a descriptive link name reduces the chance for errors (for example. 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. transmitted.2. No part of this publication may be reproduced. Shared containers have their own Category Directory and consideration should be given to a meaningful Directory Hierarchy. when sharing data with external applications (for Parallel Framework Red Book: Data Flow Job Design July 17. . Furthermore. Within a Job Sequence. stored in a retrieval system. 2.

Transformer. Sequential File. a 2. transcribed.Information Integration Solutions Center of Excellence example. etc) • The type of movement may optionally be part of the Class Word. or 4 character abbreviation should be used for the Class Word suffix.9 Stage Names DataStage assigns default names to stages as they are dragged onto the Designer canvas. In this case. Parallel Framework Red Book: Data Flow Job Design July 17. A list of frequently-used stages and their corresponding Class Word abbreviation may be found in 12.2 DataStage Naming Reference. always specify a unique link name within a particular Job or Job Sequence by including a number. reject) to reinforce the visual cues of the Designer canvas: o “Ref” for reference links (Lookup) o “Rej” for reject links (Lookup.2. establishing standardized link names makes it easier to understand results and audit counts. stored in a retrieval system. These names are based on the type of stage (object) and a unique number.4. the link name should include the link type (reference. . 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. transmitted. All rights reserved. after the subject name and subject modifier. or translated into any language in any form by any means without the written permission of IBM. 3.) • 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. Instead of using the full object name. Within a Job or Job Sequence.4. 2006 26 of 179 © 2006 IBM Information Integration Solutions. (The DataStage Designer does not require link names on different stages to be unique. stage names must be unique. Database. 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. through Job reporting). the same name may be appropriate for multiple links. No part of this publication may be reproduced. Merge.

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.

ODBC Enterprise stage read: Read the GLO. or translated into any language in any form by any means without the written permission of IBM. Oracle Enterprise stage read: Read the GLOBAL.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. transmitted. All rights reserved. . Each stage should have a short description of its function specified within the stage properties. 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. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. There are no selection criteria in the WHERE clause.GLOBAL_REST_CHAR table from jpORACLE_SERVER using the Oracle Enterprise operator. No part of this publication may be reproduced. There are no selection criteria in the WHERE clause. More complex operators or operations should have correspondingly longer and more complex explanations on this tab. 2006 30 of 179 © 2006 IBM Information Integration Solutions.RcR_GLOBAL_BUSN_CAT_TYP table from jpORACLE_SERVER using the ODBC driver. These descriptions will appear in the job documentation automatically generated from jobs and sequencers adhering to the standards in this document. transcribed.

transcribed. 2006 31 of 179 © 2006 IBM Information Integration Solutions. custom components) during job development. sequences.) that may be part of a completed application. this repository is not capable of managing nonDataStage components (for example. This stage converts null dates. 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. This stage identifies changes and drops records not matched (not updated). . Lookup stage This stage validates the input and writes rejects. Modify stage: This stage performs data conversions not requiring a transformer. Parallel Framework Red Book: Data Flow Job Design July 17. This stage validates the input and continues.Information Integration Solutions Center of Excellence Remove Duplicates stage This stage removes all but one record with duplicate BUSN_OWN_TYP_ID keys. This stage reads the GLOBAL_Lcat dataset. transmitted. which is used as write-through cache to avoid the use of Teradata. Transformer stage: This stage generates sequence numbers that have a less-than file scope. job scheduler configurations. 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. table definitions. environment files.4 Working with Source Code Control Systems DataStage’s built-in repository manages objects (jobs. stored in a retrieval system. However. This stage is cosmetic and is optimized out. routines. UNIX shell scripts. This is the target file for business qualification process rejects. 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. etc. Teradata MultiLoad stage: Load the RcR_GLOBAL_LCAT_TYP table. and to a dataset for use as write-through cache. This stage renames and/or drops columns and is NOT optimized out. 2. All rights reserved.

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

Category. There is no server-side project export facility.4. July 17. 2. In either case. The DataStage client includes Windows command-line utilities for automating the export process.3 Export to Source Code Control System The process of exporting DataStage objects to a source code control system is a straightforward process. transmitted.4. Figure 10: Manager Category browser • Choose “Export DataStage Components” from the “Export” menu. .2. stored in a retrieval system. and individual Object levels. The DataStage Manager can export at the Project.Information Integration Solutions Center of Excellence These backups can be done on a scheduled basis by an Operations support group. Assigning related objects to the same category provides a balanced level of granularity when exporting and importing objects with external source code control systems. All exports from the DataStage repository are performed on the Windows workstation.4. 2. These utilities (dsexport and dscmdexport) are documented in the DataStage Manager Guide. No part of this publication may be reproduced. establishing and following a consistent naming and categorization standard is essential to the change management process.2 Using Object Categorization Standards As discussed in Section 2. 2006 33 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. the developer should create a local backup prior to implementing any extensive changes. or by the individual DataStage developer. transcribed. or translated into any language in any form by any means without the written permission of IBM. It can be done interactively by the developer or project manager using the Manager client. as explained in this section.2 Category Hierarchy. All rights reserved. • Select the object or category in the Manager browser.

the import of objects from an external source code control system is a straightforward process.4. make sure the “Job Executables” category is checked. • Using your source code control utilities. Import can be interactive through the Manager client (as described in this section). The filename for export is specified in the “Export to file:” field at the top of the Export dialog. All rights reserved. No part of this publication may be reproduced. transmitted.Information Integration Solutions Center of Excellence NOTE: Objects cannot be exported from DataStage if they are open in Designer. or automated through command-line utilities. 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. check-in the exported . 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. Make sure all objects are saved and closed before exporting. transcribed. . stored in a retrieval system.4 Import from Source Code Control System In a similar manner.DSX file 2. the option “Selection: By category” should be specified in the “Options” tab. If you wish to include compiled Transformer objects for a selected job.

Information Integration Solutions Center of Excellence Unlike the export process. stored in a retrieval system. or using the Multi-Job Compile tool. it is possible to import the job executables from the DataStage server host using the dsjob command-line. command-line import utilities are available for both Windows workstation and DataStage server platforms. Choose “Import DataStage Components” from the “Import” menu.DSX file using Manager. This means that if necessary it will create the Job Category if it doesn't already exits.DSX file to your client workstation. If the objects were not exported with the “Job Executables”.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. 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. dsrpcd is started when the server installed.DSX file will place the object in the same DataStage category it originated from. After selecting your file. Although operating system environment variables can be set in multiple places. and should start Parallel Framework Red Book: Data Flow Job Design July 17. The Windows workstation utilities (dsimport and dscmdimport) are documented in the DataStage Manager Guide. All rights reserved. then compile the imported objects from Designer. transmitted. as documented in the DataStage Development Kit chapter of the Parallel Job Advanced Developer’s Guide. By default (in a root installation). click OK to import. transcribed. Import objects in the . Figure 12: Manager Import options • • The import of the . For test and production environments. or translated into any language in any form by any means without the written permission of IBM. 2006 35 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. 2. . • • Use the source code control system to check-out (or export) the . 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. 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.

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

2006 37 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. This setting should be disabled by default. Extremely useful for understanding how a job actually ran in the environment. Several environment variables are evaluated only for their presence in the environment (for example. Outputs EE score dump to the DataStage job log. This setting should be disabled by default. stored in a retrieval system. but part of every job design so that it can be easily enabled for debugging purposes. or may be set on an individual basis within the properties for each job. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence • causes the environment variable to be removed completely from the runtime environment. The count is per operator per partition. and Data Sets.1 Environment Variable Settings for All Jobs IBM recommends the following environment variable settings for all DataStage Enterprise Edition jobs. (see 12.1 Environment Variable Settings An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. but part of every job design so that it can be easily enabled for debugging purposes.5.4.5. This section is intended to call attention to some specific environment variables. Does not report PIDs of DataStage “phantom” processes started by Server shared containers.1. processes. These settings can be made at the project level.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. transcribed. It may be helpful to create a Job Template and include these environment variables in the parameter settings. This variable should be included in all job parameters so that it can be easily changed at runtime. 2. 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. . APT_SORT_INSERTION_CHECK_ONLY). providing detailed information about actual job flow including operators. $UNSET 2. Outputs actual runtime metadata (schema) to DataStage job log. Places entries in DataStage job log showing UNIX process ID (PID) for each process started by a job. and to document a few that are not part of the documentation. 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. transmitted.

Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.1.2 Additional Environment Variable Settings Throughout this document. No part of this publication may be reproduced.1A platforms only: On Tru64 platforms. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. . 2006 38 of 179 © 2006 IBM Information Integration Solutions. or changing the default behavior of specific Enterprise Edition stages. The environment variables mentioned in this document are summarized in Appendix D: Environment Variable Reference. you must define the environment variable $APT_IO_NOMAP On Tru64 5. 2.5.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). This setting instructs EE to use named pipes rather than shared memory for local data transport. stored in a retrieval system. All rights reserved. the environment variable $APT_PM_NO_SHARED_MEMORY should be set to 1 to work around a performance issue with shared memory MMAP operations. or translated into any language in any form by any means without the written permission of IBM. a number of environment variables will be mentioned for tuning the performance of a particular job flow. transmitted. assisting in debugging.

create re-usable components such as parallel shared containers to encapsulate frequently-used logic. user. these standard parameters and settings should be made part of a Designer Job Template. this may not be appropriate. . For example. All rights reserved. A set of standard job parameters should be used in DataStage jobs for source and target database parameters (DSN. The Multiple-Instance job property allows multiple invocations of the same job to run simultaneously. 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. or translated into any language in any form by any means without the written permission of IBM. It is for these reasons that long-running tasks are often segmented into separate jobs in an overall sequence. password. 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. No part of this publication may be reproduced. transmitted. transcribed.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. o For example. While it may be possible to construct a large.Information Integration Solutions Center of Excellence 3 Development Guidelines 3.2 Establishing Job Boundaries It is important to establish appropriate job boundaries when developing with DS/EE. complex job that satisfies given functional requirements. stored in a retrieval system. In some cases. it may be appropriate to update all dimension values before inserting new entries in a data warehousing fact table. Where possible. Create a standard directory structure outside of the DataStage project directory for source and target files. Parallel Framework Red Book: Data Flow Job Design July 17. intermediate work files. • • • 3. and so forth. etc) and directories where files are stored. 2006 39 of 179 © 2006 IBM Information Integration Solutions. processing must be restarted from the beginning of the job flow. 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. functional requirements may dictate job boundaries. o As another example. To ease re-use. But functional requirements may not be the only factor driving the size of a given DataStage job. Without these checkpoints.

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.

3. . Database stages must ALWAYS use variables for the server name. 3. Management.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. data rule violations or other reasons that would disqualify a row from processing. schema (if appropriate). Parallel Framework Red Book: Data Flow Job Design July 17. and Production Automation. Use and management of job parameters. 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. 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. any additional columns are passed through the container at runtime without the need to separate and remerge. These activities are the responsibility of job sequencers used to group jobs by some reasonable grain or by a federated scheduler. transcribed. all jobs that use a shared container must be recompiled when the container is changed. 2006 43 of 179 © 2006 IBM Information Integration Solutions.6 Parallel Shared Containers Parallel Shared Containers allow common logic to be shared across multiple jobs. Both rejects and errors will be archived and placed in a special directory for evaluation or other action by support staff. Error rows are those rows caused by unforeseen data events such as values too large for a column or text in an unsupported language. as well as standardized routines for use in Job Sequencers are discussed further in Parallel Framework Standard Practices: Administration. userid and password. All rights reserved. Using RCP. transmitted. This allows the container input and output links to contain only the columns relevant to the container processing. The presence of rejects and errors will be detected and notification sent by email to selected staff. Reject files will include those records rejected from the ETL stream due to Referential Integrity failures. stored in a retrieval system. The Usage Analysis and Multi-Job Compile tools can be used to recompile jobs that use a shared container. Because Parallel Shared Containers are inserted when a job is compiled. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. Specification of this action is the responsibility of the Business Analyst and will be published in the design document. For maximum component re-use. enable RCP at the project level and for every stage within the parallel shared container. The presence of rejects may indicate that a job has failed and prevent further processing. The exact policy for each reject is specified in the job design document. Error files will include those records from sources that fail quality tests. The presence of errors may not prevent further processing. and further.Information Integration Solutions Center of Excellence 2.

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

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

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

Only records that match the given table definition and database constraints are written. The job serial number (jpJOBSERIALNO) and a period “. and 5.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. 2. Send the reject stream to a *.7. transmitted. 3. 2 Rejects should not exist but should not stop the job.rej ECR_FACTS_DECRP_N_XformClients_dev_20060201-ETL-091504. All rights reserved. one of “rej” or “err”. 3. 3. The project phase (jpENVIRON) and a underscore “_”.7. 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. 2006 47 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. The Column Export and Transformer stages may be kept in a template Shared Container the developer will make local in each job. For example. 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. or translated into any language in any form by any means without the written permission of IBM. This section deals with both methods of handling errors. . The job name (jpJOB_NAME) and a underscore “_”. 4.5 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.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. and should be reviewed by the Data Steward.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: 1.5. transcribed.Rejects are tracked by count only. No part of this publication may be reproduced.rej file. and gathered using a Funnel stage that feeds a Sequential File stage.”. The project name (jpPROJECT_NAME) and a underscore “_”. Reject link exists. The appropriate file type. stored in a retrieval system.

transmitted. All rights reserved. stored in a retrieval system. .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. 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. Figure 13: Error Processing Components Parallel Framework Red Book: Data Flow Job Design July 17. The Transformer stage adds the required key columns. the following stages process the only errors produced by a job: The Column Export stage maps the unique columns to the single standard column.

All rights reserved. No part of this publication may be reproduced. 2006 49 of 179 © 2006 IBM Information Integration Solutions. transcribed. 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. Track*) to a single output column. transmitted. stored in a retrieval system. .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. or translated into any language in any form by any means without the written permission of IBM.

. Switch Transformer In this example. 2006 50 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. No part of this publication may be reproduced.5. OR with columns contain illegal values for some operation performed on said columns. 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. All rights reserved. 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. transcribed.Information Integration Solutions Center of Excellence Figure 15: Error Processing Transformer stage 3. The following table identifies the tagging method to be used for the previously cited operators. Connect the reject port to a Transformer stage where columns are set to specific values. attaching a nonspecific reject stream (referred to as the stealth reject stream) will gather rows from either condition to the reject stream.7. transmitted. A Transformer will reject an intact input row that cannot pass conditions specified on the output streams. A failed switch will reject an intact input row show key fails to resolve to one of the Switch output stream. 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. 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. Connect the output stream of the Transformer and Lookup stages to a Funnel stage to merge the two streams.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. or translated into any language in any form by any means without the written permission of IBM. Method Connect the reject port to a Transformer stage where those columns selected for replacement are set to specific values. Stage Lookup Description A failed lookup will reject an intact input row whose key fails to match the reference link key. 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. In either case.

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

4 Parallel Transformer stages The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. For this reason.8. 3.Information Integration Solutions Center of Excellence Server Edition Components that should be avoided within parallel job flows include: . Management.Dropping Columns July 17. always write to parallel Data Sets.3 Parallel Data Sets When writing intermediate results between DS/EE parallel jobs. without overhead for format or data type conversions. • The Copy stage should be used instead of a Transformer for simple operations including: .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. 3. it can be used at the end of a data flow o For simple jobs with only two stages. Data Sets offer parallel I/O on read and write operations. 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.BASIC Routines .8. it is best to develop a job iteratively using the Copy stage as a “placeholder”. and Production Automation. Data Sets achieve end-to-end parallelism across job boundaries by writing data in partitioned form.8. Since the Copy stage does not require an output link.Job Design placeholder between stages (unless the Force option =true.BASIC Transformers . o Unless the Force property is set to “True”. a Copy stage with a single input link and a single output link will be optimized out of the final job flow at runtime. they should not be used for long-term archive of source data. and necessary. Used in this manner. 2006 52 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. transmitted. it is important to minimize the number of transformers. in sort order. for the job control components of a DataStage Job Sequence and Before/After Job Subroutines for parallel jobs. stored in a retrieval system. Enterprise Edition will optimize this out at runtime) . All rights reserved. NOTE: Because parallel Data Sets are platform and configuration-specific.2 Copy Stage For complex data flows. and to use other stages (such as Copy) when derivations are not needed. 3. . and in Enterprise Edition native format. parallel Data Sets effectively establish restart points in the event that a job (or sequence) needs to be re-run.Renaming Columns . This is discussed in more detail in Parallel Framework Standard Practices: Administration.

1.Complex reusable logic cannot be implemented using the Transformer or . implementing complex derivation expressions using regular patterns by Lookup tables instead of using a Transformer with nested derivations. See Section 8. Parallel Framework Red Book: Data Flow Job Design July 17. The BASIC Transformer is intended as a “stop-gap” migration choice for existing Server Edition jobs containing complex routines. All rights reserved. performance should this should be tested in isolation to identify specific cause of bottlenecks. and character string trimming.1. Even then its use should be restricted and the routines should be converted as soon as possible. Instead. • Optimize the overall job flow design to combine derivations from multiple Transformers into a single Transformer stage when possible.3 Then B=”X” If A=4. No part of this publication may be reproduced. it is faster than the interpreted Filter and Switch stages. if possible. null handling. For example.5.2.6. and default type conversion can also be performed by the output mapping tab of any stage. stored in a retrieval system. . transcribed.2: Default and Explicit Type Conversions] Note that rename. • The Modify stage can be used for non-default type conversions. the derivation expression: If A=0. 2006 53 of 179 © 2006 IBM Information Integration Solutions. drop (if Runtime Column Propagation is disabled). user-defined functions and routines can expand parallel Transformer capabilities. • Because the parallel Transformer is compiled.5 BuildOp stages BuildOps should only be used when: .Information Integration Solutions Center of Excellence - Default Type Conversions [see Section 4.Existing Transformers do not meet performance requirements As always. 3. Consider.7 Then B=”C” • could also be implemented with a lookup table containing values for column A and corresponding values of column B.2: Modify Stage. The only time that Filter or Switch should be used is when the selection clauses need to be parameterized at runtime. • NEVER use the “BASIC Transformer” stage in large-volume job flows. transmitted. or translated into any language in any form by any means without the written permission of IBM.8.

LongVarBinary. No part of this publication may be reproduced. day. The following table summarizes the underlying data types of DataStage Enterprise Edition: SQL Type Internal Type Size Description Date Decimal. transmitted. with resolution of microseconds (Specify microseconds Extended option) Single field containing both date and time value with resolution to seconds. Char. uint16 int32. VarBinary Unknown. 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. transcribed. NVarChar. VarChar NChar.Information Integration Solutions Center of Excellence 4 DataStage Data Types The DataStage Designer and Manager represent column data types using SQL notation. with resolution to seconds Time of day. unit32 int64. Real Double TinyInt SmallInt Integer BigInt1 Binary. compatible with IBM packed decimal format. Parallel Framework Red Book: Data Flow Job Design July 17. (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. . Single field containing both date and time value with resolution to microseconds. Bit. uint8 int16. LongNVarChar Char. LongVarChar. 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. 2006 54 of 179 © 2006 IBM Information Integration Solutions. or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. 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. 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. All rights reserved. VarChar Time Time Timestamp Timestamp date decimal sfloat dfloat int8. Each SQL data type maps to an underlying data type in the Enterprise Edition engine. LongVarChar. Numeric Float. and year Packed decimal.

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. You can. in which case they are taken as ustrings and do require mapping. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. transcribed. specify that these data types are extended.) An Extended field appears in the columns grid. parallel jobs support two types of underlying character data types: strings and ustrings.Information Integration Solutions Center of Excellence 4. (They are specified as such by selecting the Extended check box for the column in the Edit Meta Data dialog box. however.1. ustring data represents full Unicode (UTF-16) data. 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. A blank cell indicates that no conversion is provided. Default type conversions take place across the stage output mappings of any Enterprise Edition stage. 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. transmitted.1 Strings and Ustrings If NLS is enabled on your DataStage server. . and extended Char. VarChar. The Char.1. No part of this publication may be reproduced. e = You can use a Modify or a Transformer conversion function to explicitly convert from the source field type to the destination field type. 2006 © 2006 IBM Information Integration Solutions. and LongNVarChar types relate to underlying ustring types so do not need to be explicitly extended. String data represents unmapped bytes. NVarChar. VarChar. stored in a retrieval system. 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. The NChar. or LongVarChar columns have ‘Unicode’ in this field.

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

When reading from or writing to Sequential Files or File Sets. When reading from Data Set and database sources with nullable columns. . out-of-band null representation for NULL values.Information Integration Solutions Center of Excellence IMPORTANT: When processing nullable columns in a Transformer stage. Source value propagates. Source Field not Nullable Nullable not Nullable Nullable Destination Field not Nullable Nullable Nullable not Nullable Result Source value propagates to destination. 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.1.1: Transformer NULL Handling and Reject Link. All rights reserved. the source value propagates. Source value or null propagates. July 17. stored in a retrieval system. See Section 8. care must be taken to avoid data rejects. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column. or translated into any language in any form by any means without the written permission of IBM. 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). If the source value is not null. destination value is never null. transcribed. 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. Enterprise Edition uses the internal. transmitted.

for example the Sequential File stage. it must be enabled at the project level through the Administrator client. only the columns explicitly referenced within the shared container logic need to be defined. . Using RCP. Before a DataStage developer can use RCP. or translated into any language in any form by any means without the written permission of IBM. transcribed. the remaining columns pass through at runtime.3 Runtime Column Propagation Runtime column propagation (“RCP”) allows job designs to accommodate additional columns beyond those defined by the job developer. Furthermore. Using RCP judiciously in a job design facilitates re-usable job designs based on input metadata. 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. allow their runtime schema to be parameterized further extending re-use through RCP. Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. rather than using a large number of jobs with hard-coded table definitions to perform the same tasks. a fatal error occurs. RCP facilitates re-use through parallel shared containers.Information Integration Solutions Center of Excellence If the source value is null. 2006 58 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. 4. All rights reserved. Some stages.

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

1. then a keyed partitioning method must be specified to achieve correct results. This flag is set automatically by Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. Round Robin. 2006 60 of 179 © 2006 IBM Information Integration Solutions. and job design (stage requirements and properties). Auto partitioning specifies that the Enterprise Edition engine will attempt to select the appropriate partitioning method at runtime. Auto partitioning will ensure correct results when using built-in stages. In general. Entire) and keyed (Hash) partitioning methods to produce functionally correct results and. 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. transcribed. Furthermore.1 Auto Partitioning The default partitioning method for newly-drawn links. or translated into any language in any form by any means without the written permission of IBM. Data Sets. the partitioning method may not necessarily be the most efficient from an overall job perspective. Within the Designer canvas. The “Preserve Partitioning” flag is an internal “hint” that Auto partitioning uses to attempt to preserve carefully ordered data (for example. on the output of a parallel Sort). However. stored in a retrieval system. For example. . However. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence The default partitioning method for newly-drawn links is Auto partitioning. 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. to improve performance. 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. Based on the configuration file. All rights reserved. in some cases. the ability for the Enterprise Edition engine to determine the appropriate partitioning method depends on the information available to it. if the logic defined in a Transformer stage is based on a group of related records. Auto partitioning will select between keyless (Same.

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

Since the random partition number must be calculated. or translated into any language in any form by any means without the written permission of IBM. transmitted. 3 2 1 0 . No part of this publication may be reproduced.4 Entire Partitioning Entire partitioning distributes a complete copy of the entire Data Set to each partition. Random partitioning evenly distributes rows across partitions. 2006 62 of 179 © 2006 IBM Information Integration Solutions. On clustered and Grid implementations. While in theory Random partitioning is not subject to regular data patterns that might exist in the source data. the order that rows are assigned to a particular partition will differ between job runs.3 Random Partitioning Like Round Robin. …8 7 6 5 4 3 2 1 0 ENTIRE . .2. Since optimal parallel processing occurs when all partitions have the same workload. Round Robin partitioning is useful for redistributing data that is highly skewed (there are an unequal number of rows in each partition). Same partitioning will effectively cause a downstream parallel stage to also run sequentially If you read a parallel Data Set with Same partitioning. it is rarely used in real-world data flows. but using a random assignment. . transcribed.2. similar to dealing cards: Round robin partitioning has a fairly low overhead. . 3 2 1 0 .2.1. All rights reserved. 5. the downstream stage runs with the degree of parallelism used to create the Data Set. as the complete Data Set must be distributed across the network to each node. 3 2 1 0 Parallel Framework Red Book: Data Flow Job Design July 17.1.2 Round Robin Partitioning Round Robin partitioning evenly distributes rows across partitions in a round-robin assignment.1. Entire partitioning may have a performance impact.Information Integration Solutions Center of Excellence If the upstream stage is running sequentially. Random partitioning has a slightly higher overhead than Round Robin partitioning. stored in a retrieval system. . regardless of the current $APT_CONFIG_FILE 5. 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). As a result. …8 7 6 5 4 3 2 1 0 Round Robin 6 3 0 7 4 1 8 5 2 5.

1. Assigns rows with the same values in a single integer key column to the same partition using a simple modulus calculation. Join) require processing on groups of related records. transmitted. 2006 63 of 179 © 2006 IBM Information Integration Solutions. then the resulting partitions will be of relatively equal size.3. If the source data values are evenly distributed within these key column(s). or translated into any language in any form by any means without the written permission of IBM. Remove Duplicates) or stage requirements (for example. As an example of hashing. transcribed.3 Keyed Partitioning Keyed partitioning examines the data values in one or more key columns. ensuring that records with the same values in those key column(s) are assigned to the same partition. . 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.Information Integration Solutions Center of Excellence 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 17.1. and there are a large number of unique values. Keyed partitioning is used when business rules (for example. 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.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. For DB2 Enterprise Server Edition with DPF (DB2/UDB) only – matches the internal partitioning of the specified source or target table. No part of this publication may be reproduced. All rights reserved. stored in a retrieval system. 5.

No part of this publication may be reproduced. All rights reserved. transcribed. 2006 64 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system.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. or translated into any language in any form by any means without the written permission of IBM. transmitted. .

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

Range partitioning must read the Data Set twice: once to create a Range Map file. 5. stored in a retrieval system. defeating the intention of Range partitioning.3. And. transcribed. Range partitioning assigns rows with the same values in one or more key column(s) to the same partition. No part of this publication may be reproduced. Also. if new data values are processed outside of the range of a given Range Map. 2006 66 of 179 © 2006 IBM Information Integration Solutions.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”). use Same partitioning on the input to downstream stages. if the incoming Data Set is sequential and ordered on the key column(s). A Range Map file is specific to a given parallel configuration file. typically where the incoming data values and distribution are consistent over time. Parallel Framework Red Book: Data Flow Job Design July 17. these rows will be assigned to either the first or the last partition. data is partitioned to match the internal partitioning of the target DB2 table using the DB2 partitioning method. and the second to actually partition the data within a flow using the Range Map. Using the DB2/UDB Enterprise stage. data is read in parallel from each DB2 node. Range partitioning ensures balanced workload by assigning an approximately equal number of rows to each partition.1. by default. To maintain partitioning on data read from a DB2/UDB Enterprise stage. In another scenario to avoid. it should be used if you have a single integer key column. All rights reserved. Modulus partitioning cannot be used for composite keys.3. DB2 partitioning can only be specified for target DB2/UDB Enterprise stages. In these instances. or for a non-integer key column. 5.3 Range Partitioning As a keyed partitioning method. unlike Hash and Modulus partitioning where partition skew is dependent on the actual data distribution. the Range Map file can be re-used. partition balance will be skewed. 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. transmitted. Range partitioning will result in sequential processing. . when writing data to a target DB2 database using the DB2/UDB Enterprise stage. Given a sufficient number of unique values. depending on the value. To achieve this balanced distribution. or translated into any language in any form by any means without the written permission of IBM.1. It is important to note that if the data distribution changes without recreating the Range Map.Information Integration Solutions Center of Excellence Since modulus partitioning is simpler and faster than hash.

All rights reserved. and select the “Show Instances” option. which is output to the Director job log when the environment variable APT_DUMP_SCORE is set to True.Information Integration Solutions Center of Excellence 5. in some cases. a stage’s node pool (Stage/Advanced properties) This information is detailed in the parallel job score. transcribed. as illustrated below. 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. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. transmitted. Partitions are assigned numbers.2Understanding the Parallel Job Score. right-click anywhere in the window. if specified.4. as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17.500) of rows for an optimal balanced workload. 2006 67 of 179 © 2006 IBM Information Integration Solutions. 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. This is very useful in determining the distribution across parallel partitions (skew). and each stage is processing an equal number (12. The partition number is appended to the stage name for messages written to the Director log. . 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. In this instance.2 Monitoring Partitions At runtime. Specific details on interpreting the parallel job score can be found in 12. starting at zero. No part of this publication may be reproduced. databases) c) and. the stage named “Sort_3” is running across four partitions (“x 4” next to the stage name).

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

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. stored in a retrieval system. minimize repartitioning. or translated into any language in any form by any means without the written permission of IBM. Change Capture. Remove Duplicates. persistent Data Sets can be used to retain the partitioning and sort order. 2006 69 of 179 © 2006 IBM Information Integration Solutions. Change Apply. Using the above objectives as a guide. Merge. but is not limited to: Aggregator. No part of this publication may be reproduced. transmitted. 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. transcribed.Information Integration Solutions Center of Excellence Objective 2: The partition method must match the business requirements and stage functional requirements. Objective 3: Unless partition distribution is highly skewed. 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. Join. Parallel Framework Red Book: Data Flow Job Design July 17. It may also be necessary for Transformers and BuildOps that process groups of related records. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. All rights reserved. Note that in satisfying the requirements of this second objective. 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. . This includes. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. 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. 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. and Sort stages. o Across jobs.

stored in a retrieval system. or through Auto partitioning): Parallel Framework Red Book: Data Flow Job Design July 17.4. transcribed. Additional. transmitted. Using a “standard” solution.4.2Sorting and Hashing Advanced Example. a Copy stage is used to send the detail rows to an Inner Join and an Aggregator. 2006 70 of 179 © 2006 IBM Information Integration Solutions. a Transformer is used to extract data from a single header row of an input file.4 Partitioning Examples In this section. To add aggregate columns to every detail row. This column is used as the key for a subsequent Inner Join to attach the header values to every detail row. the partitioning and sorting of this scenario can be optimized.4. 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.2 Partitioning Example 2 – Use of Entire Partitioning In this example. a new output column is defined on the header and detail links using a single constant value derivation. 5.1 Partitioning Example 1 – Optimized Partitioning The Aggregator stage only outputs key column and aggregate result columns. 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. Because the Join and Aggregator use the same partition keys and sort order. and apply Same partitioning to the downstream links. we can move the Hash partition and Sort before the Copy stage. both inputs to the Join would be Hash partitioned and sorted on this single join column (either explicitly. No part of this publication may be reproduced. more advanced partitioning and sorting examples are given in 12. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. Within the Transformer. on closer inspection. 5. The output of the Aggregator is then sent to the second input of the Join. .Information Integration Solutions Center of Excellence 5. we’ll apply the partitioning methodology defined earlier to several example job flows.

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. 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 above solution has one serious limitation.Use Round Robin partitioning on the detail input to evenly distribute rows across all partitions . In order to process a large number of detail records. the link order of the Inner Join is significant. Although functionally correct. 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: . depending on the business requirements. To optimize partitioning. there is no need to pre-sort the input to the Join. All rights reserved. 2006 71 of 179 © 2006 IBM Information Integration Solutions. transcribed. For this reason. 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). stored in a retrieval system. transmitted. it is not always the most efficient solution. so we will revisit this in the Sorting discussion. the link order in this example should be set so that the single header row is assigned to the Right input.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.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. consider that the single header row is really a form of reference data. resulting in sequential processing. Remembering that the degree of parallel operation is limited by the number of distinct values.

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. Like partitioning methods.Information Integration Solutions Center of Excellence If defined in reverse of this order. 2006 72 of 179 © 2006 IBM Information Integration Solutions. transmitted.3: Minimizing Runtime Processes and Resource Requirements). the order of rows in an Auto collector is undefined.5. stored in a retrieval system.5 Collector Types Collectors combine parallel partitions of an input Data Set (single link) into a single input stream to a stage running sequentially. or translated into any language in any form by any means without the written permission of IBM. 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. there is one further detail in this example. when the previous stage is running in parallel as shown on the right: Figure 33: Specifying Collector method 5. there is a specialized example where the Round Robin collector may be appropriate. No part of this publication may be reproduced. transcribed. The Round Robin collector is generally slower than an Auto collector because it must wait for a row to appear in a particular partition. 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. and may vary between job runs on the same Data Set. . However. the collector method is defined in the stage Input/Partitioning properties for any stage running sequentially. Auto is the default collector method.5. the Join will attempt to read all detail rows from the right input (since they have the same key column value) into memory.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.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. For this reason. the detail rows in the Left input will buffer to disk to prevent a deadlock. 5. (See Section 12. All rights reserved. 5. For advanced users.

5. This is because Round Robin collector reads from partitions using the same partition order that a Round Robin partitioner assigns rows to parallel partitions.4 Sort Merge Collector If the input Data Set is sorted in parallel. as long as the Data Set has not been repartitioned or reduced. 5. through aggregation). 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. and these should be the same columns.5. in the same order. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. 5. No part of this publication may be reproduced. use Sort Merge collector to produce a single. as used to sort the input Data Set in parallel. Row order is undefined for non-key columns. 2006 73 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17. . stored in a retrieval system. the Sort Merge collector will generate a sequential stream of rows in globally sorted order.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. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. Ordered collectors are generally only useful if the input Data Set has been Sorted and Range partitioned on the same key column(s). transcribed.3 Ordered Collector An Ordered collector reads all rows from the first partition.5. All rights reserved.6 Collecting Methodology Given the options for collecting data into a sequential stream. transmitted. 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. The Sort Merge collector requires one or more key column(s) to be defined. use Auto partitioning (the default) b) When the input Data Set has been sorted in parallel. or translated into any language in any form by any means without the written permission of IBM. then reads all rows from the next partition until all rows in the Data Set have been collected. In this scenario. an Ordered collector will generate a sequential stream in sort order.

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

2006 75 of 179 © 2006 IBM Information Integration Solutions. Using Remove Duplicates. we want to select the most recent order for a given customer. sequential ordered result is needed. For example. To satisfy these requirements we could:  Partition on CustID to group related records  Sort on OrderDate in Descending order  Remove Duplicates on CustID. stored in a retrieval system.2 Sorting and Hashing Advanced Example provides a more detailed discussion and example of partitioning and sorting. . Let’s say that an input Data Set consists of order history based on CustID and Order Date. 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). or translated into any language in any form by any means without the written permission of IBM. in general it is best to use a two step process: . transcribed.2 Complete (Total) Sort If a single. No part of this publication may be reproduced. 6. transmitted.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. All rights reserved. with Duplicate To Retain=First 12. 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.4.partition and parallel Sort on key column(s) Parallel Framework Red Book: Data Flow Job Design July 17.

stored in a retrieval system. The Link sort offers fewer options. partitioning. but is easier to maintain in a DataStage job.sort direction: Ascending or Descending . the “Perform Sort” option is checked. the standalone Sort stage is used to specify the Sort Key mode for partial sorts. use the Link sort unless a specific option is needed on the stand-alone Stage.3. No part of this publication may be reproduced. In general. When specifying key column(s) for partitioning. or both . . or translated into any language in any form by any means without the written permission of IBM. both methods use the same internal sort package (the tsort operator). transcribed. when specifying a keyed partitioning method.1 Link Sort Sorting on a link is specified on the Input/Partitioning stage options. transmitted.case sensitivity (strings) . The stand-alone sort offers more options. 2006 76 of 179 © 2006 IBM Information Integration Solutions.Sorting character set: ASCII (default) or EBCDIC (strings) Parallel Framework Red Book: Data Flow Job Design July 17. but as a separate stage makes job maintenance slightly more complicated.key column usage: sorting. (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. Within the Designer canvas. Most often. although the Enterprise Edition engine may insert a sort if required). ordered result set This is similar to the way parallel database engines perform their parallel sort operations. 6. as there are fewer stages on the design canvas. 6.3 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).Information Integration Solutions Center of Excellence - use a Sort Merge collector on these same key column(s) to generate a sequential. By default. All rights reserved. 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: .

This “subsort” uses significantly less disk space and CPU resource. Parallel Framework Red Book: Data Flow Job Design July 17. and thus a stable sort is generally slower than a non-stable sort for the same input Data Set and sort keys. and can often be performed in memory (depending on the size of the new subsort groups). which is significantly faster than a “UNIX” sort. 6. 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. All rights reserved. NOTE: The Sort Utility option is an artifact of previous releases.4 Stable Sort Stable sorts preserve the order of non-key columns within each sort group. . It is used when resorting a sub-grouping of a previously sorted input Data Set. Specifically. This requires some additional overhead in the sort algorithm. For this reason. No part of this publication may be reproduced. 2006 77 of 179 © 2006 IBM Information Integration Solutions. 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. Always specify “DataStage” Sort Utility. stored in a retrieval system. transcribed.2 Sort Stage The standalone Sort stage offers more options than the sort on a link. transmitted. instead of performing a complete Sort. the key column property “Sort Key Mode” is a particularly powerful feature and a significant performance optimization.5 Sub-Sorts Within the standalone Sort stage. 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. or translated into any language in any form by any means without the written permission of IBM. - 6.3.

Enterprise Edition inserts sorts before any stage that requires matched key values or ordered groupings of (Join. 2006 78 of 179 © 2006 IBM Information Integration Solutions. all key columns must still be defined in the Sort stage. keys with “Don’t Sort (Previously Sorted)” property must be at the top of the list.2 Understanding the Parallel Job Score) can be used to identify automatically-inserted sorts. transmitted.p2] node4[op2. as shown in the following example: Figure 38: Sort Key Mode property To successfully perform a subsort. If the input data does not match the key column definition for a subsort. The parallel job score (see 12. key={value=FirstName}}(0)) on nodes ( node1[op2. transcribed. Remove Duplicates. DataStage Enterprise Edition inserts sort operators as necessary to ensure correct results. 6. the job will abort.6 Automatically-Inserted Sorts By default. without gaps between them. stored in a retrieval system. All rights reserved. No part of this publication may be reproduced. Sorts are only inserted automatically when the flow developer has not explicitly defined an input sort. while new sort keys are specified with the “Sort” key mode property.p1] node3[op2. Parallel Framework Red Book: Data Flow Job Design July 17.p0] node2[op2.p3] )} Typically. Re-used sort keys are specified with the “Don’t Sort (Previously Sorted)” property.4. the key column order for these keys must match the key columns and order defined in the previously-sorted input Data Set. Merge. as shown in this score fragment: op1[4p] {(parallel inserted tsort operator {key={value=LastName}. . And. or translated into any language in any form by any means without the written permission of IBM. Sort Aggregator).Information Integration Solutions Center of Excellence To resort based on a sub-grouping.

sequential ordered result set use a parallel Sort and a Sort Merge collector 6. memory. There are two ways to prevent Enterprise Edition from inserting an un-necessary sort: a) Insert an upstream Sort stage on each link. stored in a retrieval system. 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. the environment variable APT_TSORT_STRESS_BLOCKSIZE can be Parallel Framework Red Book: Data Flow Job Design July 17. transmitted. the environment variable $APT_SORT_INSERTION_CHECK_ONLY should be set to prevent Enterprise Edition from inserting un-necessary sorts before the Join stage. 2006 79 of 179 © 2006 IBM Information Integration Solutions.7 Sort Methodology Using the rules and behavior outlined in the previous section. and disk resources. To perform a sort. inserted sorts can be a significant performance impact if they are not necessary. This value can be changed for each standalone Sort stage using the “Restrict Memory Usage” option (the minimum is 1MB/partition). On a global basis. .4: Partitioning Examples. No part of this publication may be reproduced. By default. Revisiting the partitioning examples in Section 5. 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.Information Integration Solutions Center of Excellence While ensuring correct results. All rights reserved. aborting the job if data is not in the required sort order. rows in the input Data Set are read into a memory buffer on each partition. 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.8 Tuning Sort Sort is a particularly expensive task within DataStage Enterprise Edition which requires CPU. This will verify sort order but not actually perform a sort. transcribed. Create Key Change Column. or translated into any language in any form by any means without the written permission of IBM. Create Cluster Key Change Column. each sort uses 20MB of memory per partition for its memory buffer. 6. If the sort operation can be performed in memory (as is often the case with a subsort) then no disk I/O is performed.

for all sort operators (link and standalone). Parallel Framework Red Book: Data Flow Job Design July 17. Having a greater number of scratch disks for each node allows the sort to spread I/O across multiple file systems. in MB.scratch disks defined in the current configuration file (APT_CONFIG_FILE) in the “sort” named disk pool .Information Integration Solutions Center of Excellence use to specify the size of the memory buffer. 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 default directory specified by the environment variable TMPDIR . then results are temporarily spooled to disk in the following order: . overriding any per-sort specifications.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. transmitted. If the input Data Set cannot fit into the sort memory buffer.scratch disks defined in the current configuration file default disk pool . . transcribed. stored in a retrieval system. 2006 80 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.

as no overhead is needed to translate data to the internal DS/EE representation. such as mainframe sources with COBOL copybook file definitions. performance penalty of conversion. Need to share information with external applications. If data is Parallel Framework Red Book: Data Flow Job Design July 17.1 Which File Stage to Use DataStage/EE offers various stages for reading from and writing to files. are summarized below: File Stage Recommended Usage Read and write standard files in a single format.2 Data Set Usage Parallel Data Sets are the persistent (on-disk) representation of the in-memory data structures of DS/EE. Data Sets can only be read from and written to using a DataStage parallel job. can write in parallel (generates multiple segment files). and any limitations. Data Sets provide maximum performance for reading and writing data from disk.Information Integration Solutions Center of Excellence 7 File Stage Usage 7. Slightly higher overhead than Data Set. In general. transcribed. Can only be used as reference link on a Lookup stage. As such. However. or translated into any language in any form by any means without the written permission of IBM. 7. does not support hierarchical data files. (Requires SAS connectivity license for DataStage. Sequential File Need to read source data in complex (hierarchical) format. Some stages (parallel Data Set) support “Append” to add new records to an existing file. Intermediate storage between DataStage parallel jobs. Cannot write in parallel. Need to share data with an external Parallel SAS application. transmitted. Requires Parallel SAS. can only be read from / written to by DS/EE or Parallel SAS. but this is not recommended as it imposes risks for failure recovery. All rights reserved. using the internal format of the parallel engine. Can only be read from and written to by DataStage parallel jobs or orchadmin command. Complex Flat File Data Set File Set SAS Parallel Data Set Lookup File Set No DS/EE file stage supports “update” of existing records. stored in a retrieval system. . No part of this publication may be reproduced. Limitations Cannot write to a single file in parallel. Data Sets store data in partitioned form. Recommendations for when to use a particular stage. 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. performance penalty of format conversion. 2006 81 of 179 © 2006 IBM Information Integration Solutions.

A better option for writing to a set of Sequential Files in parallel is to use the FileSet stage. set environment variable $APT_IMPORT_PATTERN_USES_FILESET Read Method: Specific Files. and then pass this to a Column Import stage to parse the file in parallel.3.3 Separating I/O from Column Import If the Sequential File input cannot be read in parallel. 2006 82 of 179 © 2006 IBM Information Integration Solutions.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.file may be either fixed or variable-width Read Method: Specific Files. transmitted. stored in a retrieval system. The formatting and column properties of the Column Import stage match those of the Sequential File stage. No part of this publication may be reproduced. Read From Multiple Nodes option is set to Yes useful for cluster and Grid configurations . Parallel Framework Red Book: Data Flow Job Design July 17.3. . in parallel. only one file specified. Unlike the Complex Flat File stage.3. define a single large string column for the non-parallel Sequential File read. The FileSet stage will write in parallel. This will create a single header file (in text format) and corresponding data files. Readers Per Node option greater than 1 useful for SMP configurations . All rights reserved.Information Integration Solutions Center of Excellence 7. the degree of parallelism of the write will correspond to the number of file names specified.file may only be fixed-width Note that when reading in parallel. or translated into any language in any form by any means without the written permission of IBM. 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. more than one file specified each file specified within a single Sequential File stage must be of the same format Read Method: File Pattern. As shown in the following Job fragment. performance can still be improved by separating the file I/O from the column parsing operation. 7. transcribed. using the format options specified in the FileSet stage. In these instances. the Sequential File stage can only read and write data that is in flattened (row/column) format. input row order is not maintained across readers.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.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. 7.

• 7.4 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. the buffers are always flushed to disk.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. or translated into any language in any form by any means without the written permission of IBM. . 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. All rights reserved. including Sequential File import / export format properties is documented in the Orchestrate Record Schema manual.3. It is important to use ROUNDROBIN partitioning (or other partitioning appropriate to downstream components) to evenly distribute the data in the flow. the in-band (value) must be explicitly defined in the extended column attributes for each Nullable column.3. SAME will read the entire file into a single partition.3. instead of statically through Table Definitions. 7. No part of this publication may be reproduced.3. 7. 2006 83 of 179 © 2006 IBM Information Integration Solutions. When a job completes successfully. or by using a File Pattern). each file’s data is read into a separate partition.7 Reading and Writing Nullable Columns When reading from or writing to Sequential Files or File Sets. Note that this document is required. transcribed. but there is a small performance penalty associated with increased I/O. It is also important to remember that this setting will apply to all Sequential File stages in the data flow. Using the Schema File option allows the format of the source file to be specified at runtime. stored in a retrieval system. Setting this value to a low number (such as 1) is useful for realtime applications. 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. as shown below: Parallel Framework Red Book: Data Flow Job Design July 17. transmitted.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. the Sequential File (export operator) stage buffers its writes to optimize performance. making the entire downstream flow run sequentially (unless it is later repartitioned). The format of the Schema File. When multiple files are read by a single Sequential File stage (using multiple files.5 Sequential File (Export) Buffering By default. 7.

By default. Decimal. All import and export properties are listed in the Import/Export Properties chapter of the Orchestrate Operators Reference. To display each field value. 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. • If a field is nullable. Double-click on the column number in the grid dialog to set this column property.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. Decimal. . No part of this publication may be reproduced. you must define the null field value and length in the Nullable section of the column property.9 Reading Bounded-Length VARCHAR Columns Care must be taken when reading delimited. bounded-length Varchar columns (Varchars with the length option set).3. transmitted. When writing fixed-length files from variable-length fields (eg. Integer. use the print_field import property. stored in a retrieval system. All rights reserved. Double-click on the column number in the grid dialog to set this column property. • • 7.3. 2006 84 of 179 © 2006 IBM Information Integration Solutions. Use caution when specifying this option as it can generate an enormous amount of detail in the job log. or translated into any language in any form by any means without the written permission of IBM. Varchar). Varchar). Integer. the field width and pad string column properties must be set to match the fixed-width of the output column.Information Integration Solutions Center of Excellence Figure 40: Extended Column Metadata (Nullable properties) 7. transcribed. the field width column property must be set to match the fixed-width of the input column. if the source file has fields with values longer than the maximum Parallel Framework Red Book: Data Flow Job Design July 17.

the stage allows you to write data to one or more complex flat files. but you can configure the stage to execute sequentially if it is only reading one file with a single reader. Increasing this size may improve performance. (This is an error in the DataStage documentation. 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. A complex flat file may contain one or more GROUPs. 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. . the environment variables $APT_IMPORT_BUFFER_SIZE and $APT_EXPORT_BUFFER_SIZE can be used to improve I/O performance. in some disk array configurations. including MVS datasets with QSAM and VSAM files. or OCCURS clauses. REDEFINES. Complex Flat File source stages execute in parallel mode when they are used to read multiple files.) When used as a target. these extra characters will be silently truncated. It does not write to MVS datasets.Information Integration Solutions Center of Excellence Varchar length. stored in a retrieval system. transmitted. 7. All rights reserved. 7. the stage allows you to read data from one or more complex flat files. 2006 85 of 179 © 2006 IBM Information Integration Solutions. 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. No part of this publication may be reproduced.3. These settings specify the size of the read (import) and write (export) buffer size in Kbytes. NOTE: The Complex Flat File stage cannot read from sources with OCCURS DEPENDING ON clauses.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. with a default of 128 (128K).10 Tuning Sequential File Performance On heavily-loaded file servers or some RAID/SAN array configurations. transcribed. When used as a source. Finally.

2006 86 of 179 © 2006 IBM Information Integration Solutions.y] decimal[x+y. . trailing separate. 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.y] decimal[x+y.y] or string[x+y] decimal[x+y.y] decimal[x+y. native binary binary.y] or string[x+y] decimal[x+y.y] decimal[x+y. 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. transcribed.y] decimal[x+y. All rights reserved.y] sfloat dfloat ustring[n] ustring[max=n] subrec Internal Options packed packed zoned zoned. stored in a retrieval system.1 CFF Stage Data Type Mapping When you work with mainframe data using the CFF stage. native binary binary. trailing zoned. 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. trailing zoned.4. leading separate. No part of this publication may be reproduced. leading Parallel Framework Red Book: Data Flow Job Design July 17. native binary binary. or translated into any language in any form by any means without the written permission of IBM. transmitted.Information Integration Solutions Center of Excellence 7. native binary binary. native binary binary.

stored in a retrieval system. When rows are rejected by a Transformer. it is important to minimize the number of transformers. For example. Even if the target column in an output derivation allows nullable results. @ROWNUM is Parallel Framework Red Book: Data Flow Job Design July 17. 8.1.col Note that if an incoming column is only used in an output column mapping. or translated into any language in any form by any means without the written permission of IBM. .Information Integration Solutions Center of Excellence 8 Transformation Languages 8. entries are placed in the Director job log. To create a Transformer reject link in Designer. 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.1 Transformer NULL Handling and Reject Link When evaluating expressions for output derivations or link constraints. All rights reserved.4: Parallel Transformer stages for guidelines on Transformer stage usage. For this reason. 8.1. if you intend to use a nullable column within a Transformer derivation or output link constraint.2 Parallel Transformer System Variables The system variable @ROWNUM behaves differently in the Enterprise Edition Transformer stage than in the Server Edition Transformer. This makes it easy to identify reject conditions (by row counts). 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. For this reason. the Transformer will reject (through the reject link indicated by a dashed line) any row that has a NULL value used in the expression. Always include reject links in a parallel Transformer. and to use other stages (such as Copy) when derivations are not needed. Because the DS/EE Transformer runs in parallel. transmitted. transcribed. the Transformer will reject the row instead of sending it to the output link(s).1 Transformer Stage The DataStage Enterprise Edition parallel Transformer stage generates “C” code which is then compiled into a parallel component. No part of this publication may be reproduced.8. by definition. 2006 87 of 179 © 2006 IBM Information Integration Solutions. undefined.col) Then “” Else link. the Transformer will allow this row to be sent to the output link without being rejected. the following stage variable expression would convert a null value to a specific empty string: If ISNULL(link.

Examples: 1. stored in a retrieval system. Examples: 1. Within the link constraints dialog box.Information Integration Solutions Center of Excellence assigned to incoming rows for each partition. or translated into any language in any form by any means without the written permission of IBM.5 -> -2 Parallel Framework Red Book: Data Flow Job Design July 17. By default. the system variables @NUMPARTITIONS and @PARTITIONNUM should be used. the PadString function uses the length of the source type. 1. there are times when internal decimal variables need to be generated in order to perform the evaluation. When generating a sequence of numbers in parallel.3 Transformer Derivation Evaluation Output derivations are evaluated before any type conversions on the assignment. By default.5 Transformer Decimal Arithmetic When decimal data is evaluated by the Transformer stage. 1). Since the Transformer will abort the entire job flow immediately. All rights reserved. it is possible that valid rows will not have been flushed from Sequential File (export) buffers.4 -> -2 round_inf Rounds or truncates towards nearest representable value. -1.6 ->1. or committed to database tables. not the target. For example. Therefore. For example. Examples: 1. or performing parallel derivations. . -1. transmitted. 2006 88 of 179 © 2006 IBM Information Integration Solutions. No part of this publication may be reproduced. Create a new output link that will handle rows that match the abort rule. 8. the incoming column must be type VarChar before it is evaluated in the Transformer. 8. transcribed. Thus.3. internal decimal results are rounded to the nearest applicable value. TrimLeadingTrailing(string) works only if string is a VarChar field. up to a maximum precision of 255 and scale of 125. apply the abort rule to this output link.1. -1. these internal decimal variables will have a precision and scale of 38 and 10.4 -> 1.5-> 2.4 -> 2.1. If more precision is required. -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.5: Sequential File (Export) Buffering). and set the “Abort After Rows” count to the number of rows allowed before the job should be aborted (for example. 8. breaking ties by rounding positive values toward positive infinity and negative values toward negative infinity.4 Conditionally Aborting Jobs The Transformer can be used to conditionally abort a job when incoming data matches a specific rule. the environment variables APT_DECIMAL_INTERM_PRECISION and APT_DECIMAL_INTERM_SCALE can be set to the desired range. It is important to set the database commit parameters or adjust the Sequential File buffer settings (see Section 7. it is important to make sure the type conversion is done before a row reaches the Transformer.6 -> -1 floor Rounds towards negative infinity.4 -> -1.1.

2006 89 of 179 © 2006 IBM Information Integration Solutions. transmitted. In this case. if $APT_DECIMAL_INTERM_SCALE is smaller than the results of the internal calculation. Examples: 1. transcribed. 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.1.56 -> 1. the substring is evaluated just once for every input row. suppose multiple columns in output links want to use the same substring of an input column. unless the derivation is empty For each output link: Evaluate the link constraint.3] = “001”) THEN . The evaluation sequence is as follows: Evaluate each stage variable initial value For each input row to process: Evaluate each stage variable derivation value. By doing this. .col[1. stored in a retrieval system. the evaluation of the substring of DSLINK1... 8. Such constructs are: • Where the same part of an expression is used in multiple column derivations For example. the stage variable definition would be: Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced.6 Optimizing Transformer Expressions and Stage Variables In order to write efficient Transformer stage derivations. it can be seen that there are certain constructs that would be inefficient to include in output column derivations. This can be made more efficient by moving the substring calculation into a stage variable. For example.col[1. In this case. then the following test may appear in a number of output columns derivations: IF (DSLINK1. -1. the output links are also evaluated in the order in which they are displayed. round or truncate to the scale size. as they would be evaluated once for every output column that uses them. it is useful to understand what items get evaluated and when.3] is evaluated for each column that uses it. or translated into any language in any form by any means without the written permission of IBM.56 ->-1. All rights reserved.5. Similarly.Information Integration Solutions Center of Excellence trunc_zero Discard any fractional digits to the right of the rightmost fractional digit supported regardless of sign.5. From this sequence.

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

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

and justify defaults to begin.2 Modify and String Trim The function string_trim has been added to Modify. 8. To convert from an in-band null to an out-of-band null. By default. . value can be a string. 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. string. 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. or translated into any language in any form by any means without the written permission of IBM. For a numeric field value can be a numeric value.dataType is its optional data type. date.destField is the destination field’s name. stored in a retrieval system. use it if you are also converting types. transmitted.sourceField is the source field’s name.2. direction defaults to end.value) where: . with the following syntax: stringField=string_trim[character. Justify has no affect when the target string has variable length. 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. 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. . these characters are retained when the fixed-length string is then converted back to a variable-length string. By default.value) where: . All rights reserved. the syntax is: destField[:dataType] = handle_null (sourceField. . this is NULL. The character argument is the character to remove. .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. 2006 92 of 179 © 2006 IBM Information Integration Solutions. time. and timestamp fields.value is the value of the source field when it is null. direction. The value of the direction and justify arguments can be either begin or end. The destField is converted from an Orchestrate out-of-band null to a value of the field’s data type.destField is the destination field’s name.value is the value you wish to represent a null in the output.Information Integration Solutions Center of Excellence 8.sourceField is the source field’s name . transcribed.2. No part of this publication may be reproduced. for decimal. . . use it if you are also converting types.

2006 93 of 179 © 2006 IBM Information Integration Solutions. end. All rights reserved.Information Integration Solutions Center of Excellence name:string = string_trim[NULL. begin](name) The following example removes all trailing Z characters from color. transmitted. . stored in a retrieval system. begin](color) Parallel Framework Red Book: Data Flow Job Design July 17. and left-justifies the resulting hue fixed-length string: hue:string[10] = string_trim[‘Z’. 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.

All rights reserved. Left Outer. and ODBC Enterprise stages) to scenarios where the number of input rows is significantly smaller (for example. all rows on an outer link (for example. the JOIN or MERGE stage should be used. If un-matched rows must be captured or logged.1. it is best to use a Column Generator to add an ‘indicator’ column. For this reason. 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. In most cases.Information Integration Solutions Center of Excellence 9 Combining Data 9. the default value for an Integer is zero. No part of this publication may be reproduced. During an Outer 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.1 Lookup vs. an OUTER join operation must be performed. In an OUTER join scenario. This is most easily done by inserting a Copy stage and mapping a column from NON-NULLABLE to NULLABLE. For example. Oracle Enterprise. or translated into any language in any form by any means without the written permission of IBM. and the default value for a Char is a string of padchar characters equal to the length of the Char column. If the Data Sets are larger than available memory resources. It is best to test both the Sparse and Normal to see which actually performs best. This isolates your match/no-match logic from any changes in the metadata. or both links in the case of Full Outer) are output regardless of match on key values. Limit the use of database Sparse Lookups (available in the DB2 Enterprise. . 9.2 Capturing Unmatched Records from a Join The Join stage does not provide reject handling for unmatched records (such as in an InnerJoin scenario). care must be taken to change the column properties to allow NULL values before the Join. with a constant value. to each of the inner links and test that column for the constant after you have performed the join. and to retest if the relative volumes of data change dramatically. Join). Join vs. This is also handy with Lookups that have multiple reference links. stored in a retrieval system. Right Outer. 2006 94 of 179 © 2006 IBM Information Integration Solutions. transcribed. Each lookup reference requires a contiguous block of shared memory. A Transformer stage can be used to test for NULL values in unmatched columns. transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. the default value for a Varchar is an empty string (“”). (see Section 10. when a match does not occur. then Enterprise Edition inserts “default” values based on the data type. Sparse Lookups may also be appropriate for exception-based processing when the number of exceptions is a small fraction of the main input data.7: Database Sparse Lookup vs. 1:100 or more) than the number of reference rows.

the Sort Aggregator requires presorted data. An exception to this is financial calculations which should be done in decimal to preserve appropriate precision. To aggregate in decimal precision.3. Unlike the Hash Aggregator.3. 9. stored in a retrieval system.2 Aggregation Data Type By default. 2006 95 of 179 © 2006 IBM Information Integration Solutions.3 The Aggregator Stage 9. To perform a total aggregation. All rights reserved. the output data type of a parallel Aggregator stage calculation or recalculation column is floating point (Double). but only maintains the calculations for the current group in memory.3. 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. or translated into any language in any form by any means without the written permission of IBM. transcribed. .3 Performing Total Aggregations The Aggregator counts and calculates based on distinct key value groupings. 9. No part of this publication may be reproduced. 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. which maintains the results of each keycolumn value/aggregation pair in memory.1 Aggregation Method By default. Parallel Framework Red Book: Data Flow Job Design July 17. 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.Information Integration Solutions Center of Excellence 9. set the optional property “Aggregations/Default to Decimal Output” within the Aggregator stage. 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. The Sort Aggregation Method should be used when the number of key values is unknown or very large. You can also specify that the result of an individual calculation or recalculation is decimal by using the optional “Decimal Output” sub-property. transmitted.

Teradata API) are visible in the default DataStage Designer palette. You may need to customize the palette to add hidden stages. For some databases (DB2.1 Database stage types DataStage Enterprise Edition offers database connectivity through native parallel and plug-in stage types. and Teradata).Information Integration Solutions Center of Excellence 10 Database Stage Guidelines 10. transcribed. Subsequent sections provide database-specific tips and guidelines. 10. Oracle. 2006 96 of 179 © 2006 IBM Information Integration Solutions. transmitted.1. 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. 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. All rights reserved. Informix. .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. 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.

Because of their tight integration with database technologies. they should only be used when it is not possible to use a native parallel stage. operating system. the native parallel stages often have more stringent connectivity requirements than plug-in stages. . 10. This allows Enterprise Edition to match return columns by name. transmitted. not name. “Wire Protocol” ODBC Drivers generally do not require database client software to be installed on the server platform. and version.1. Native parallel stages always pre-query the database for actual runtime metadata (column names. On UNIX. and features it is best to use the native parallel database stages in a job design if connectivity requirements can be satisfied.1. not position in the stage Table Definitions. 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). However. Because plug-in stage types cannot read in parallel.1. For example. plug-in database stages match columns by order. DataStage Enterprise Edition bundles OEM versions of ODBC drivers from DataDirect. scalability. 10. or translated into any language in any form by any means without the written permission of IBM.1.1. No part of this publication may be reproduced. attributes). Parallel Framework Red Book: Data Flow Job Design July 17. Unlike the database-specific parallel stages. From a design perspective. types. 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.Information Integration Solutions Center of Excellence 10. specific guidelines of when to use various stage types are provided in the database-specific topics in this section. Because there are exceptions to this rule (especially with Teradata).1 Native Parallel database stages In general. 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. Furthermore. and cannot span multiple servers in a clustered or Grid configuration. the ODBC Enterprise stage cannot interface with database-specific parallel load technologies. stored in a retrieval system.3 Plug-In database stages Plug-in stage types are intended to provide connectivity to database configurations not offered by the native parallel stages. ODBC also provides an increased level of “data virtualization” which can be useful when sources and targets (or deployment platforms) can change. for maximum parallel performance. transcribed.2 ODBC Enterprise stage In general. All rights reserved. the DB2/UDB Enterprise stage is only compatible with DB2 Enterprise Server Edition with DPF on the same UNIX platform as the DataStage server. native database components (such as the Oracle Enterprise stage) are preferable to ODBC connectivity if both are supported on the database platform. the DataDirect ODBC Driver Manager is also included.1.

2.columns of the database row correspond to columns of a DS/EE record .2 Database Metadata 10.rows of the database result set correspond to records of a DS/EE Data Set . stored in a retrieval system. This allows the database stages to match return values by column name instead of position. 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. 2006 98 of 179 © 2006 IBM Information Integration Solutions. data types. Informix Enterprise or Oracle Enterprise stages. the DS/EE native parallel database stages always “pre-query” the database source or target to determine the actual metadata (column names.1.2 Metadata Import When using the native parallel DB2 Enterprise.2.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 . regardless of the table definitions assigned by the DataStage developer. 10. nullability) and partitioning scheme (in some cases) of the source or target table.1 Runtime metadata At runtime. 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. transmitted. Database-specific data type mapping tables are included in the following sections. . No part of this publication may be reproduced.1.both DS/EE and relational databases support null values. use orchdbutil to import metadata to avoid type conversion issues. or translated into any language in any form by any means without the written permission of IBM. However. All rights reserved.Information Integration Solutions Center of Excellence 10. For each native parallel database stage: . This utility is available as a server command line utility and within Designer and Manager using “Import Orchestrate Schema Definitions”. and selecting “Import from Database Table” option in the wizard as illustrated below: Parallel Framework Red Book: Data Flow Job Design July 17.1. transcribed.

For example. No part of this publication may be reproduced. it is important to use SQL aliases to explicitly name the calculated columns so that they can be referenced within the DataStage job. For “Table” read method. However. instead of using an unqualified “Table” or SQL “SELECT *” read. As a command. it will be easier to use the corresponding orchdbutil command-line utility from the DataStage server machine. you will see a window with the result columns and data.1. For “Auto-Generated” SQL. 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. When importing a large number of tables. 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. 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.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.3 Optimizing Select Lists For best performance and optimal memory usage. the DataStage Designer will automatically populate the select list based on the stage’s output column definition. 10. 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.2. All rights reserved. The alias name(s) should then be added to the Table Definition within DataStage. orchdbutil can be scripted to automate the process of importing a large number of tables.1. there may be cases where user-defined functions or logic need to be executed on the database server. always specify the “Select List” subproperty. 10. . Test the connection using View Data button. transmitted.1.4 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. transcribed. If the connection is successful. similar to the illustration on the right: Parallel Framework Red Book: Data Flow Job Design July 17. 2006 99 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system.

it is not a good idea to let DataStage generate target tables unless they are used for temporary storage. the results should be landed to a parallel Data Set. Subsequent job(s) should read this Data Set and populate the target table using the appropriate database stage and write method.6 Database OPEN and CLOSE Commands The native parallel database stages provide options for specifying OPEN and CLOSE commands. constraints.Information Integration Solutions Center of Excellence Figure 43: Sample View Data Output If the connection fails. it is important to separate the transformation process from the database write (Load or Upsert) operation.5 Designing for Restart To enable restart of high-volume jobs. As another example.1. 10. transcribed. There are limited capabilities to specify Create table options in the stage. the OPEN command can be used to create a target table. the OPEN command could be used to create a temporary table.1. and doing so may violate data-management (DBA) policies. 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. 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. As a further optimization. including databasespecific options (tablespace. logging. . All rights reserved. After transformation. depending on data volume) can be used to identify existing rows before they are inserted into the target table. Figure 44: View Additional Error Detail 10. a Lookup stage (or Join stage. Parallel Framework Red Book: Data Flow Job Design July 17. 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. stored in a retrieval system. 2006 100 of 179 © 2006 IBM Information Integration Solutions. In general. and you will be prompted to view additional detail. For example. and the CLOSE command could be used to select all rows from the temporary table and insert into a final target table. an error message may appear. transmitted. etc) not possible with the “Create” option. OPEN and CLOSE are not offered by plugin database stages.

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

transmitted. or when the source includes a large number of database tables. For maximum scalability and parallel performance. In this scenario. 2006 102 of 179 © 2006 IBM Information Integration Solutions. Avoid the use of database stored procedures (for example. especially when the join columns are indexed. All rights reserved. the efficiency of the Enterprise Edition Sort and Join stages can be significantly faster than an equivalent SQL query. it can still be beneficial to use database filters (WHERE clause) if appropriate. Oracle PL/SQL) on a per-row basis within a high-volume data flow. 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. A join that reduces the result set significantly is also often appropriate to do in the database. stored in a retrieval system. transcribed. . When combining data from very large tables. • • Parallel Framework Red Book: Data Flow Job Design July 17.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. it is best to implement business rules using native parallel DataStage components.

The DB2/UDB Enterprise stage requires DB2 Enterprise Server Edition on UNIX with Data Partitioning Facility (DPF) option.1. 3 A patched version of the ODBC Enterprise stage allowing parallel read is available from IBM IIS Support for some platforms. 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. or translated into any language in any form by any means without the written permission of IBM. upsert. If you use the stage in this way.) Furthermore. transcribed.1 DB2 Stage Types DataStage Enterprise Edition provides access to DB2 databases using one of 5 stages. lookup. . the DB2 hardware/UNIX/software platform must match the hardware/software platform of the DataStage ETL server.2 DB2 Guidelines 10. No part of this publication may be reproduced. 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. the “client authentication” configuration of a remote instance is not recommended. stored in a retrieval system. 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.Information Integration Solutions Center of Excellence 10. 2006 103 of 179 © 2006 IBM Information Integration Solutions.2.2. consult the DataStage documentation (DataStage Parallel Job Developers Guide. For this reason. transmitted. and load capabilities to parallel DB2 databases on UNIX using the native parallel DB2/UDB Enterprise stage. Parallel Framework Red Book: Data Flow Job Design July 17. Check with IBM IIS Support for availability. this was also called “DB2 EEE”. All rights reserved.1 DB2/UDB Enterprise stage Enterprise Edition provides native parallel read. This will only work when the authentication mode of the database on the remote instance is set to “client authentication”. (Before DB2 v8. and DataStage Plug-In guides) 10. 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.

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

DB2INSTANCE. • User. to identify the instance name of the DB2 server. stored in a retrieval system.cfg file can July 17. Set this to the DB2 client’s alias database name for the remote DB2 server database. • Database. Otherwise use the DB2 environment variable. Set this to the DB2 client instance name. 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. DataStage assumes you require remote connection. 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. Optionally set this to the instance name of the DB2 server. • Client Alias DB Name. it is possible to connect to more than one DB2 instance within a single job. Enter the user name for connecting to DB2.3 Configuring DB2 Multiple Instances in One DataStage Job Although it is not officially supported.2. 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. transmitted. 2006 105 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. you will need to set the following properties on the DB2/UDB Enterprise stage in your parallel job: • Client Instance Name. Single stream . Enter the password for connecting to DB2. This is required only if the client’s alias is different from the actual name of the remote server database. . Once this variable is set. Optionally set this to the remote server database name. If you set this property. Since a db2nodes. 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.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. or translated into any language in any form by any means without the written permission of IBM. 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. it will try to use it for each of the connections in the job. 10. transcribed. 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. • Password. • Server. All rights reserved. Otherwise use the environment variables $APT_DBNAME or $APT_DB2DBDFT to identify the database. No part of this publication may be reproduced.Information Integration Solutions Center of Excellence Figure 45: DB2/UDB Enterprise stage properties For connection to a remote DB2/UDB instance.

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

Once the table definition is loaded. Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. transcribed. stored in a retrieval system. and you should use them if entering SQL in the job yourself. The DB2/UDB Enterprise stage converts DB2 data types to Enterprise Edition data types.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. 2006 107 of 179 © 2006 IBM Information Integration Solutions. the internal column names are displayed rather than the original DB2 names both in table definitions and in the Data Browser. DB2 Data Type CHAR(n) CHARACTER VARYING(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. transmitted. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. They are also used in derivations and expressions. avoid hand editing (this minimizes the risk of mistakes or confusion). 10.6 DB2/UDB Enterprise stage Data Type Mapping The DB2 database schema to be accessed must NOT have any columns with User Defined Types (UDTs). the result is a time field. .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.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. DECIMAL[p.2. as shown in the following table. the result is a timestamp field. All rights reserved. Alternatively. or translated into any language in any form by any means without the written permission of IBM. The original names are used in generated SQL statements. decimal[p.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour. examine the DDL for each schema to be accessed.r) NVARCHAR(n. No part of this publication may be reproduced. however.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n.

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

These values are in turn used to read the SYSTABLEPART table to retrieve the number of partitions. 2006 109 of 179 © 2006 IBM Information Integration Solutions. P. Finally the SYSKEYS and SYSCOLUMNS tables are read using the index name to get the associated column metadata (name and type). Table T is in tablespace TS and TS is partitioned into 3 partitions on Col1 (limits: F. 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. T) and Col2 (10. This information determines the number of db2read operators that the conductor builds into the score and the queries that they execute. No part of this publication may be reproduced. 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. This is illustrated in Figure 47. functions within the db2read operator are used to read the DB2 SYSTABLES table to retrieve the tablespace and database name for the table. transmitted. stored in a retrieval system.Information Integration Solutions Center of Excellence When accessing a DB2 table using the Table read method. and the partition limit key value(s). . Parallel Framework Red Book: Data Flow Job Design July 17. as illustrated in Figure 46: Figure 46: DB2 read on DataStage/USS For example. 30. partitioning index name(s). the number of operators do not have to match to the number of partitions. transcribed. and is controlled by the number of nodes in the configuration file. All rights reserved. 40).

or translated into any language in any form by any means without the written permission of IBM. stored in a retrieval system. An example of an in-memory Normal Lookup is shown in Figure 48. 2006 110 of 179 © 2006 IBM Information Integration Solutions. . All rights reserved. transmitted. No part of this publication may be reproduced. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed. 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.Information Integration Solutions Center of Excellence Figure 47: DB2 write on DataStage/USS On DataStage/USS Edition.

Figure 49: DB2 Sparse Lookup on DataStage/USS Finally. All rights reserved. 2006 111 of 179 © 2006 IBM Information Integration Solutions. we need to add a special resource statement in our configuration file to specify the MVS dataset name to use. . Since each of these queries must go through the DB2 coordinator node we can effectively ignore the level of parallelism specified for the table. transmitted. where each lookup operator is issuing an SQL to DB2 for every row it processes. When the conductor creates the score. 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. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. Contrast the Normal Lookup with the way a Sparse Lookup is done as shown in Figure 49. 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. Since there is no sequential file stage associated with this MVS load file.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. The DB2 LOAD utility is designed to run from JCL only. 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. using the DB2 load utility in USS is different from non-z/OS environments. In order to invoke it from a DataStage/USS job. nor can it be read in from a USS HFS file. Parallel Framework Red Book: Data Flow Job Design July 17. we call a DB2 stored procedure called DSNUTILS. No part of this publication may be reproduced. The LOAD utility has a second limitation in that data cannot be piped into it. transcribed.

stored in a retrieval system. No part of this publication may be reproduced. .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. 2006 112 of 179 © 2006 IBM Information Integration Solutions. transmitted. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. transcribed.

r) DATE DATETIME Enterprise Edition Data Type string[n] string[max=n] date date. If the DATETIME starts with a year component. 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. No part of this publication may be reproduced.rows of the database result set correspond to records of a DS/EE Data Set .2 Informix Enterprise stage Data Type Mapping Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions.1 Informix Enterprise Stage Column Names For each Informix Enterprise stage: .columns of the database row correspond to columns of a DS/EE record . DECIMAL[p. The Informix Enterprise stage converts Informix data types to Enterprise Edition 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 . Informix Data Type CHAR(n) CHARACTER VARYING(n.3 Informix Database Guidelines 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.Information Integration Solutions Center of Excellence 10. and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. 2006 113 of 179 © 2006 IBM Information Integration Solutions. time or timestamp with corresponding fractional precision for time: If the DATETIME starts with a year component and ends with a month.r) REAL SERIAL SMALLFLOAT SMALLINT VARCHAR(n) If the DATETIME starts with an hour.s] DOUBLE-PRECISION FLOAT INTEGER MONEY NCHAR(n.3. as shown in the following table. or translated into any language in any form by any means without the written permission of IBM.r) NVARCHAR(n.s] where p is the precision and s is the scale The maximum precision is 32. transmitted.both DS/EE and Informix support null values. the result is a time field. the result is a timestamp field. .3. the result is a date field. All rights reserved. transcribed. decimal[p.

names are translated exactly except when the external data source column name contains a character that DataStage does not support. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10.2 ODBC Enterprise stage Data Type Mapping ODBC data sources are not supported by the orcdbutil utility. transmitted.4.both DS/EE and ODBC support null values.columns of the database row correspond to columns of a DS/EE record . two underscore characters replace the unsupported character . stored in a retrieval system. 2006 114 of 179 © 2006 IBM Information Integration Solutions.rows of the database result set correspond to records of a DS/EE Data Set .4 ODBC Enterprise Guidelines 10.1 ODBC Enterprise Stage Column Names For each ODBC Enterprise stage: .s] decimal[p. or translated into any language in any form by any means without the written permission of IBM.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] where p is the precision and s is the scale decimal[p.s] string[36] int32 int8 [0 or 1] decimal[p. Enterprise Edition will return an error and abort your job. 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. transcribed. If you attempt to read a record larger than 32K. It is important to verify the correct ODBC to Enterprise Edition data mapping. Parallel Framework Red Book: Data Flow Job Design July 17. No part of this publication may be reproduced. All rights reserved.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.4. In that case.Information Integration Solutions Center of Excellence 10.

2006 115 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. . and will generate an error at runtime Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.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. No part of this publication may be reproduced. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. transmitted.

the ODBC Enterprise stage does not support parallel read4. it may be possible to write to a target database in parallel using the ODBC Enterprise stage. . Depending on the target database. All rights reserved. transmitted. stored in a retrieval system. since this capability is not provided by the ODBC API. and the table configuration (row or page level lock mode if available).Information Integration Solutions Center of Excellence 10. Parallel Framework Red Book: Data Flow Job Design July 17.4. 4 On some platforms. 2006 116 of 179 © 2006 IBM Information Integration Solutions.3 Reading ODBC Sources Unlike other native parallel database stages. Parallel reads through ODBC match the degree of parallelism in the $APT_CONFIG_FILE. transcribed. a patch may be available through IBM IIS Support to support parallel reads through ODBC. 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.

or translated into any language in any form by any means without the written permission of IBM. Enterprise Edition will return an error and abort your job. Examples include: . and will generate an error at runtime 10. stored in a retrieval system. The underlying Oracle table does not have to be partitioned for parallel read within Enterprise Edition.5 Oracle Database Guidelines 10. In that case. It is important to note that certain types of queries cannot run in parallel.Information Integration Solutions Center of Excellence 10.5.columns of the database row correspond to columns of a DS/EE record .5. which are not heavily typed. If you attempt to read a record larger than 32K.names are translated exactly except when the Oracle source column name contains a character that DataStage does not support. the Oracle Enterprise stage reads sequentially from its source table or query. All rights reserved.queries containing a GROUP BY clause that are also hash partitioned on the same field Parallel Framework Red Book: Data Flow Job Design July 17.2 Oracle Enterprise stage Data Type Mapping Oracle Table Definitions should be imported into DataStage using orchdbutil to ensure accurate Table Definitions. 2006 117 of 179 © 2006 IBM Information Integration Solutions. .5.rows of the database result set correspond to records of a DS/EE Data Set .3 Reading from Oracle in Parallel By default.both DS/EE and Oracle support null values. transcribed.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. IMPORTANT: Oracle data types that are not listed in the above table cannot be used in the Oracle Enterprise stage.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 .1 Oracle Enterprise Stage Column Names For each Oracle Enterprise stage: . This is particularly important for Oracle databases. two underscore characters replace the unsupported character . Enterprise Edition maps Oracle data types based on the rules given in the following table: Oracle Data Type CHAR(n) DATE NUMBER NUMBER[p. transmitted.s] RAW(n) VARCHAR(n) Enterprise Edition Data Type string[n] or ustring[n] a fixed-length string with length = n timestamp decimal[38. Setting the partition table option to the specified table will enable parallel extracts from an Oracle source. and a null value in a database column is stored as an out-of-band NULL value in the DS/EE column 10. No part of this publication may be reproduced.10] int32 if precision(p) < 11 and scale (s) = 0 decimal[p.

In order to automatically generate the SQL required by the Upsert method. rebuild). • Setting the environment variable $APT_ORACLE_LOAD_OPTIONS to “OPTIONS (DIRECT=TRUE. the Oracle load will be done sequentially. No part of this publication may be reproduced. 2006 118 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. or translated into any language in any form by any means without the written permission of IBM. In this instance. Enterprise Edition uses the Parallel Direct Path Load method.4 Oracle Load Options When writing to an Oracle table (using Write Method = Load). Parallel Framework Red Book: Data Flow Job Design July 17. .5. When using this method.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. transmitted. the key column(s) must be identified using the check boxes in the column grid. All rights reserved. PARALLEL=FALSE) also allows loading of indexed tables without index maintenance. 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. The Upsert Write Method can be used to insert rows into a target Oracle table without bypassing indexes or constraints. transcribed.

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.

Indirectly. All rights reserved. Teradata Enterprise stage uses 32K buffers. (Note that 64K buffers must be enabled at the Teradata server level). and will generate an error at runtime. Parallel Framework Red Book: Data Flow Job Design July 17.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.password=password[. transmitted.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. stored in a retrieval system.RequestedSessions=nn] where SesionsPerPlayer and RequestedSessions are optional connection parameters that are required when accessing large Teradata databases. To specify a Teradata password that contains special characters.6 Specifying Teradata Passwords with Special Characters Teradata permits passwords with special characters and symbols. but this can be set to a value between 1 and the database vprocs. 2006 122 of 179 © 2006 IBM Information Integration Solutions. In that case SessionsPerPlayer should be increased. the password must be surrounded by an “escaped” single quote as shown. By default.7 Teradata Enterprise Settings Within the Teradata Enterprise stage.7. RequestedSessions equals the maximum number of available sessions on the Teradata instance. 10. Setting the SessionsPerPlayer too low on a large system can result in so many players that the job fails due to insufficient resources. No part of this publication may be reproduced.SessionsPerPlayer=nn][. and/or RequestedSessions should be decreased. transcribed. By default. Aggregates and most arithmetic operators are not allowed in the SELECT clause of a Teradata Enterprise stage. . 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. 10. and hence the number of UNIX processes and overall system resource requirements of the DataStage job. or translated into any language in any form by any means without the written permission of IBM.7.7. this determines the number of DataStage players. The SessionsPerPlayer option determines the number of connections each DataStage EE player opens to Teradata. the DB Options property specifies the connection string and connection properties in the form: user=username. where pa$$ is the example password: \’pa$$\’ 10.

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

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

or translated into any language in any form by any means without the written permission of IBM. look at row counts. transmitted. and to manage source or target Parallel Data Sets. 2006 125 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. stored in a retrieval system. enable both $OSH_PRINT_SCHEMAS and $DS_PX_DEBUG . . the log entry is sometimes preceded by a warning condition. it is possible for $OSH_PRINT_SCHEMAS to generate a log entry that is too large for DataStage Director to store or display.    NOTE: For large jobs. These may indicate an underlying logic problem or unexpected data type conversion. Set the environment variable $DS_PX_DEBUG to capture copies of the job score. Use the Data Set Management tool (available in the Tools menu of DataStage Designer or DataStage Manager) to examine the schema.1: Environment Variable Settings and the DataStage Parallel Job Advanced Developers Guide. No part of this publication may be reproduced. this variable determines the number of seconds the conductor node will wait for a successful startup from each section leader. it may not be possible to remove all warning messages generated by the EE engine. When a fatal error occurs. This directory will be created in the “Debugging” sub-directory of the Project home directory on the DataStage server. All rights reserved. To capture the full schema output in these cases.Information Integration Solutions Center of Excellence $APT_PM_NODE_TIMEOUT [seconds] For heavily loaded MPP or clustered environments. July 17. or promote a job from development into test or production. tune. generated osh. The general process for debugging a job is:  Check the Director job log for warnings. 11.  Examine the score dump (placed in the Director log when $APT_DUMP_SCORE is enabled).  Enable the Job Monitoring Environment Variables detailed in Section 2. In some instances. The default is 30 seconds. But all warnings should be examined and understood. Use $OSH_PRINT_SCHEMAS to verify that the job’s runtime schemas matches what the job developer expected in the design-time column definitions. transcribed. All fatal and warning messages should be addressed before attempting to debug.5.3 How to Isolate and Debug a Parallel Job There are a number of tools available to debug DataStage Enterprise Edition jobs. This will place entries in the Director log with the actual runtime schema for every link using Enterprise Edition internal data types. and internal Enterprise Edition log messages in a directory corresponding to the job name.

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. the generated OSH tab will appear in the Job Properties dialog box: Parallel Framework Red Book: Data Flow Job Design July 17. All rights reserved. and to understand internally what is running. jobs are compiled into OSH (Orchestrate SHell) scripts that are used to execute the given job design at runtime. use the UNIX command od –xc –Ax o To display the number of lines and characters in a specified ASCII text file. 11.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). this count may be incorrect. or translated into any language in any form by any means without the written permission of IBM.4 Viewing the Generated OSH Within Designer. NOTE: The wc command counts UNIX line delimiters. transmitted. It is also not useful for files of non-delimited fixed-length record format. 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. so if the file has any binary columns. transcribed. 2006 126 of 179 © 2006 IBM Information Integration Solutions. stored in a retrieval system. . No part of this publication may be reproduced. It is useful to examine the generated OSH for debugging purposes.

where data is repartitioned and how it is repartitioned. stored in a retrieval system. that is. All stages in this flow are running on one node. transcribed. including how composite operators and shared containers break down. what degree of parallelism each operator runs with. have been inserted by EE. The Peek and Sort stages are combined. they have been optimized into the same process. The job runs 3 processes on 2 nodes. A score dump includes a variety of information about a flow. which operators. the first task is to examine the score dump which is generated when you set APT_DUMP_SCORE=TRUE in your environment. Sort (tsort) and Peek. . or translated into any language in any form by any means without the written permission of IBM. which has a hash partitioner that partitions on key field a. ##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. All rights reserved.Information Integration Solutions Center of Excellence Figure 52: Generated OSH in Designer Job Properties 11. Also available is some information about where data may be buffered. and exactly which nodes each operator runs on. if any. transmitted.5 Interpreting the Parallel Job Score When attempting to understand an Enterprise Edition flow. The following score dump shows a flow with a single Data Set. It shows three stages: Generator. No part of this publication may be reproduced. 2006 127 of 179 © 2006 IBM Information Integration Solutions.

are the computation-intensive stages shared evenly across all nodes? More details on interpreting the parallel job score can be found in 12.Information Integration Solutions Center of Excellence lemond.torrent.com[op0. Parallel Framework Red Book: Data Flow Job Design July 17.2 Understanding the Parallel Job Score. . No part of this publication may be reproduced. All rights reserved.com[op1.4. transcribed. stored in a retrieval system. there are three areas to investigate: • Are there sequential stages? • Is needless repartitioning occurring? • In a cluster or Grid.com[op1.torrent. transmitted.p1] )} In a score dump.torrent.p0] )} op1[2p] {(parallel APT_CombinedOperatorController: (tsort) (peek) )on nodes ( lemond. 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.

a) Use Parallel Data Sets to land intermediate result between parallel jobs.1 How to Design a Job for Optimal Performance Overall job design can be the most significant factor in data flow performance. 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. stored in a retrieval system. 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). They maintain reference data in DS/EE internal format. facilitating end-to-end parallelism across job boundaries. If you need to share information with external applications. • Data Sets can only be read by other DS/EE parallel jobs (or the orchadmin command line utility). b) Remove unneeded columns as early as possible within the data flow. Parallel Framework Red Book: Data Flow Job Design July 17. proceeds through integration and volume testing. No part of this publication may be reproduced.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. File Sets facilitate parallel I/O at the expense of exporting to a specified file format. pre-indexed. This section provides tips for designing a job for optimal performance. transcribed. This section outlines performance-related tips that can be followed when building a parallel data flow using DataStage Enterprise Edition. 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. in the DS/EE native internal format. 12. . • When reading from database sources. • Parallel Data Sets retain data partitioning and sort order. All rights reserved. Lookup File Sets can only be used on reference links to a Lookup stage. Performance tuning and optimization is an iterative process that begins at job design and unit tests. transmitted. 2006 129 of 179 © 2006 IBM Information Integration Solutions. • Lookup File Sets can be used to store reference data used in subsequent jobs. However. or translated into any language in any form by any means without the written permission of IBM. and continues throughout an application’s production lifecycle.

and to simply re-sort on a differing set of secondary/tertiary/etc. Note. f) Minimize the number of partitioners in a job. use orchdbutil to ensure that the designtime metadata matches the actual runtime metadata (especially with Oracle databases). All rights reserved. Copy. When business rules dictate otherwise and the data volume is large and sufficiently skewed. For data type conversions. there is greater control over the efficiency of code. that unless dynamic (parameterized) conditions are required. a Transformer is always faster than a Filter or Switch stage. o When working with database sources and targets. Modify) may be more appropriate. Parallel Framework Red Book: Data Flow Job Design July 17. Use BuildOps only when existing Transformers do not meet performance requirements. • When possible. • • Avoid using the BASIC Transformer. • There are limited scenarios when the memory overhead of handling large Varchar columns would dictate the use of unbounded strings. ensure data is as close to evenly distributed as possible. External user-defined functions can expand the capabilities of the parallel Transformer. or when complex reusable logic is required. at the expense of ease of development (and more skilled developer requirements). transmitted. stored in a retrieval system. 2006 130 of 179 © 2006 IBM Information Integration Solutions. especially in large-volume data flows. Because BuildOps are built in C++. other stages (for example. 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. however. Unbounded strings (Varchar’s without a maximum length) can have a significant negative performance impact on a job flow. 32K) maximum length that are rarely populated  Varchar columns of a large maximum length with highly varying data sizes d) Avoid type conversions.  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. For example:  Varchar columns of a large (for example. It is usually possible to choose a smaller partition-key set. repartition to a more balanced distribution as soon as possible to improve performance of downstream stages. renaming and removing columns. if possible. keys. .Information Integration Solutions Center of Excellence c) Always specify a maximum length for Varchar columns.

.2 Understanding Operator Combination At runtime. by using “SAME” partitioning. When reading from these Data Sets. there is nothing to be gained from pipeline partitioning5. Combined operators are generally adjacent to each other in a data flow. your results should be tested in your environment. stored in a retrieval system. When composing the score. Parallel Framework Red Book: Data Flow Job Design July 17. o When writing to parallel Data Sets.Information Integration Solutions Center of Excellence • • Know your data. The purpose behind operator combination is to reduce the overhead associated with an increased process count. In these configurations. and should only be used if there is a need to maintain an implied (i. 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. 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. Remember that SAME maintains the degree of parallelism of the upstream operator. All rights reserved. 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. transcribed. specifying the “don’t sort. The default setting is 20MB per partition. if possible. 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. No part of this publication may be reproduced. sort order and partitioning are preserved. or translated into any language in any form by any means without the written permission of IBM. 5 One exception to this guideline is when operator combination generates too few processes to keep the processors busy. transmitted. o The stable sort option is much more expensive than non-stable sorts. partitioning.e. even those that have the “Restrict Memory Usage” option set. previously sorted” option for those key columns in the Sort stage will reduce the cost of sorting and take greater advantage of pipeline parallelism. try to maintain this sorting.  If data has already been partitioned and sorted on a set of key columns. As with any example. not explicitly stated in the sort keys) row order. 12. and groupings. Use SAME partitioning carefully. disabling operator combination allows CPU activity to be spread across multiple processors instead of being constrained to a single processor. If two processes are interdependent (one processes the other’s output) and they are both CPU-bound or I/O-bound. 2006 131 of 179 © 2006 IBM Information Integration Solutions. 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). In addition.

Information Integration Solutions Center of Excellence However. as shown in the following illustration: Figure 53: CPU-bound combined process in Job Monitor Parallel Framework Red Book: Data Flow Job Design July 17.0 versions of DS/EE. combination can be set on a per-stage basis (on the Stage/Advanced tab) The job score identifies what components are combined.2 Understanding the Parallel Job Score” in this document. There are 2 ways to affect operator combination: o The environment variable APT_DISABLE_COMBINATION. it is best to let DSEE decide what to combine and what to leave uncombined. 2006 132 of 179 © 2006 IBM Information Integration Solutions. this is only recommended on pre 7. the assumptions used by the Enterprise Edition optimizer to determine which stages can be combined may not always be the most efficient. Enterprise Edition is ‘greedy’ .a. or globally. When deciding which operators to include in a particular combined operator (a. stored in a retrieval system.4. All rights reserved. if the “%CPU” column is displayed in a Job Monitor window in Director. combined stages are indicated by parenthesis surrounding the % CPU. (For information on interpreting a job score dump.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’.k. Combined Operator Controller). see 12. tuning combination might yield additional performance benefits. However. disables ALL combination in the entire data flow.) In addition. It is for this reason that combination can be enabled or disabled on a per-stage basis. No part of this publication may be reproduced. o Within Designer. . transmitted. 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. when other performance tuning measures have been applied and still greater performance is needed. or translated into any language in any form by any means without the written permission of IBM. transcribed.

it is good to separate I/O heavy operators (Sequential File. All rights reserved.Information Integration Solutions Center of Excellence Choosing which operators to disallow combination for is as much art as science. in general. in the above job design. Disabling combination with the Transformer enables pipeline partitioning. such as Join. No part of this publication may be reproduced. 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. However. the I/O-intensive FileSet is combined with a CPU-intensive Transformer. Change Capture. however. In fact. stored in a retrieval system.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. it might be a good idea to set the sequential file to be non-combinable. For example. DS/EE executes a given job across the resources defined in a the specified configuration file. 2006 133 of 179 © 2006 IBM Information Integration Solutions. . a single-node configuration file is often appropriate to minimize job startup time and resource requirements without significantly impacting overall performance. and improves performance. Toward that end. or translated into any language in any form by any means without the written permission of IBM. when it is appropriate to minimize the resource requirements for a given scenario. as shown in this subsequent Job Monitor for the same job: Figure 54: Throughput in Job Monitor after disabling combination 12. if you have several transformers and database operators combined with an output Sequential File. Full Sorts.) from CPU-heavy operators (Transformer. etc. 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. Lookup. Merge. transcribed. This will prevent IO requests from waiting on CPU to become available and viceversa.). etc. transmitted. Funnel) Parallel Framework Red Book: Data Flow Job Design July 17. There are times.

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

2006 135 of 179 © 2006 IBM Information Integration Solutions.4.Information Integration Solutions Center of Excellence    o Specifies the maximum allowable block size for transferring data between players. No part of this publication may be reproduced. 12. and APT_AUTO_TRANSPORT_BLOCK_SIZE are used only with fixed-length records. must be at least 2 NOTE: The environment variables APT_MIN/MAX_TRANSPORT_BLOCK_SIZE. Here is an example job fragment: Figure 55: Fork-Join example Parallel Framework Red Book: Data Flow Job Design July 17. . APT_LATENCY_COEFFICIENT. cannot be greater than 1048576. APT_AUTO_TRANSPORT_BLOCK_SIZE and APT_MMIN_TRANSPORT_BLOCK_SIZE APT_AUTO_TRANSPORT_BLOCK_SIZE o If set. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. This variable is only meaningful when used in combination with APT_LATENCY_COEFFICIENT. Default is 1048576. Cannot be less than 8192. “Deadlock Prevention” comes into play anytime there is a Fork-Join structure in a job.2 Deadlock Prevention Buffering The other type of buffering. transmitted. stored in a retrieval system. transcribed. 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.

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

Values greater than 1 indicate that the buffer July 17. Join cannot be ‘stuck’ waiting to read from either of its inputs. stored in a retrieval system. BufferOps will also be placed on the input partitions to any sequential stage that is fed by a parallel stage. 2006 137 of 179  Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. Default 1 MB. “The Parallel Engine Configuration File” of the DataStage Manager guide). BufferOps will allocate 3MB of memory each (remember that this is per operator. transmitted. APT_BUFFER_FREE_RUN o Maximum capacity of the buffer operator before it starts to offer resistance to incoming flow.Information Integration Solutions Center of Excellence BufferOp1 Aggregator 1 Transformer Join Aggregator2 BufferOp 2 (Here the arrows now represent data-flow. as these same types of circular dependencies can result from partition-wise Fork-Joins. transcribed. thus breaking the circular dependency and guaranteeing no deadlock will occur. 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. not dependency.) Since BufferOp is always ready to read or write. 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. 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. May not exceed 2/3 of APT_BUFFER_MAXIMUM_MEMORY. All rights reserved. By default. . it may be necessary to increase the default buffer size (APT_BUFFER_MAXIMUM_MEMORY) to hold more rows in memory. per partition). No part of this publication may be reproduced. TIP: For very wide rows.

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

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

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

  DataStage Data Types Be aware of the mapping between DataStage (SQL) data types and the internal DS/EE data types. Further details are provided in Section 3. Parallel Framework Red Book: Data Flow Job Design July 17. . minimizing overall run time. Objective 2: The partition method must match the business requirements and stage functional requirements. 2006 141 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. Server Shared Containers) within a parallel job. 3. as outlined in Section 4:DataStage Data Types. transmitted. 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. Standardized Error Handling routines should be followed to capture errors and rejects.  Always use parallel Data Sets for intermediate storage between jobs. and to facilitate default type conversions. transcribed. This ensures that the processing workload is evenly balanced. Component Usage As discussed in Section 3. while minimizing overhead. the following guidelines should be followed when constructing parallel jobs in DS/EE:  Never use Server Edition components (BASIC Transformer. stored in a retrieval system.  Use BuildOp stages only when logic cannot be implemented in the parallel Transformer. Partitioning Data Given the numerous options for keyless and keyed partitioning. No part of this publication may be reproduced.  Use the parallel Transformer stage (not the BASIC Transformer) instead of the Filter or Switch stages. 4. or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Parallel Shared Containers should be used to encapsulate frequently-used logic. 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.7:Error and Reject Record Handling. BASIC Routines are appropriate only for job control sequences. using RCP to maximize re-use.  Use the Copy stage as a placeholder for iterative design. 5.8: Component Usage. Leverage default type conversions using the Copy stage or across the Output mapping tab of other stages.

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. Remove Duplicates. Merge. Change Capture. but is not limited to: Aggregator. Change Apply. Using the above objectives as a guide. Collecting Data Given the options for collecting data into a sequential stream. especially in cluster or Grid configurations Repartitioning data in a cluster or Grid configuration incurs the overhead of network transport. and Sort stages. the following guidelines form a methodology for choosing the appropriate collector type: a) When output order does not matter. . Note that in satisfying the requirements of this second objective. 6. 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. This may require re-examining key column usage within stages and re-ordering stages within a flow (if business requirements permit). All rights reserved. transmitted. 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. or translated into any language in any form by any means without the written permission of IBM. minimize repartitioning. 2006 142 of 179 © 2006 IBM Information Integration Solutions. transcribed. It may also be necessary for Transformers and BuildOps that process groups of related records.Information Integration Solutions Center of Excellence This includes. use Auto partitioning (the default) Parallel Framework Red Book: Data Flow Job Design July 17. Join. No part of this publication may be reproduced. 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. examine up-stream partitioning and sort order and attempt to preserve for down-stream processing. it may not be possible to choose a partitioning method that gives close to an equal number of rows in each partition. stored in a retrieval system. persistent Data Sets can be used to retain the partitioning and sort order. Objective 3: Unless partition distribution is highly skewed. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. Across jobs.

Information Integration Solutions Center of Excellence b) When the input Data Set has been sorted in parallel. as long as the Data Set has not been repartitioned or reduced. 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. use the Join or Merge stage.  Parallel Framework Red Book: Data Flow Job Design July 17.1: Transformer NULL Handling and Reject Link. Create Key Change Column. 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. Join 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. Sorting Using the rules and behavior outlined in Section 6: Sorting. Create Cluster Key Change Column.1: Lookup vs. Merge. 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. Further details on Partitioning methods can be found in Section 5: Partitioning and Collecting. 7. 2006 143 of 179 © 2006 IBM Information Integration Solutions. The Lookup stage is most appropriate when reference data is small enough to fit into available memory. No part of this publication may be reproduced. globally sorted stream of rows o When the input Data Set has been sorted in parallel and Range partitioned. o Always place a reject link on a parallel Transformer to capture / audit possible rejects. transcribed. stored in a retrieval system. . If the Data Sets are larger than available memory resources. or translated into any language in any form by any means without the written permission of IBM. use Sort Merge collector to produce a single. sequential ordered result set use a parallel Sort and a Sort Merge collector 8.  Stage-Specific Guidelines As discussed in Section 8.1. See Section 9.

use orchdbutil to properly import design metadata. No part of this publication may be reproduced. 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. When using Oracle. or translated into any language in any form by any means without the written permission of IBM. use a SQL where clause to limit the number of rows sent to a DataStage job. DB2. which may indicate an underlying problem or data type conversion issue. Even if the source data is not nullable. or when exception processing. . Parallel Framework Red Book: Data Flow Job Design July 17. For maximum scalability and parallel performance. Avoid the use of database stored procedures on a per-row basis within a high-volume data flow.1.   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. A Sort method Aggregator should be used when the number of distinct key values is large or unknown. All rights reserved. use the native parallel database stages for maximum performance and scalability. (See Section 9. 10. All warnings and failures should be addressed (and removed if possible) before deploying a DS/EE job. 9. Be particularly careful to observe the nullability properties for input links to any form of Outer Join. or Informix databases. stored in a retrieval system.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.  Database Stage Guidelines Where possible. transmitted. transcribed. as discussed in Section 10. Check the Director log for warnings. 2006 144 of 179 © 2006 IBM Information Integration Solutions. If possible. it is best to implement business rules natively using DataStage parallel components.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. Use Hash method Aggregators only when the number of distinct key column values is small. the non-key columns must be defined as nullable in the Join stage input in order to identify unmatched records.2Capturing Unmatched Records from a Join).

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

transmitted. No part of this publication may be reproduced. 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. File. 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. . or translated into any language in any form by any means without the written permission of IBM.Information Integration Solutions Center of Excellence Appendix B: DataStage Naming Reference Every name should be based on a three-part concept: Subject. 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. stored in a retrieval system. Subject Modifier. Class Word where the following frequently-used Class Words describe the object type. transcribed.

No part of this publication may be reproduced. transmitted. . or translated into any language in any form by any means without the written permission of IBM. transcribed.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. stored in a retrieval system. 2006 147 of 179 © 2006 IBM Information Integration Solutions. All rights reserved.

2006 148 of 179 © 2006 IBM Information Integration Solutions. At runtime. Enterprise Edition uses the given job design and configuration file to compose a job score which details the processes created. transmitted. not the actual job. transcribed. The first score is from the license operator. 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. and can be ignored. 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. multiple operators are combined within a single operating system process to improve performance and optimize resource requirements. stored in a retrieval system. Parallel Framework Red Book: Data Flow Job Design July 17. 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 job score is output to the DataStage Director log.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. 1. It is recommended that this setting be enabled by default at the project level. Similar to the way a parallel database optimizer builds a query plan. as the job score offers invaluable data for debugging and performance tuning. 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. The second score entry is the actual job score. and interconnects (Data Sets) between them. Viewing the Job Score When the environment variable APT_DUMP_SCORE is set. As shown in the illustration below. and the overhead to capture the score is negligible. . degree of parallelism and node (server) assignments. All rights reserved.

July 17.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. “node4”). 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. 2006 149 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. This includes: o Sequential or Parallel operation. . two running in parallel across 4 nodes. there are 3 operators. stored in a retrieval system. • Operators: starts with the words “It has n operators:” The second section details actual operators created to execute the job flow. including persistent (on disk) and virtual (in memory. Parallel Job Score Components The Enterprise Edition parallel job score is divided into two sections. links between stages). (in this example: “node1”. for a total of 9 operating system process. transcribed. transmitted. there are two virtual Data Sets. The actual node names correspond to node names in the parallel configuration file. “node3”. one running sequentially. In this example. All rights reserved. “node2”. In this 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. and the degree of parallelism per operator o Node assignment for each operator.

collector type with the second. corresponding to the lower section of the job score. for persistent Data Sets) July 17. transcribed. For example. In the above example. with each Data Set identified by its number (starting at zero). Within the Data Set definition. 3. while operators are identified by their operator number and name. 2006 150 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. the first Data Set is identified as “ds0”. 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. followed by a notation to indicate the type of partitioning or collecting (if any). No part of this publication may be reproduced. 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 1 (op1) is running in parallel with 4 degrees of parallelism [4p]. operator zero (op0) is running sequentially. with 1 degree of parallelism [1p]. followed by the downstream consumer. stored in a retrieval system.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. All rights reserved. Persistent Data Sets are identified by their Data Set name. the upstream producer is identified first. not SAME) Parallel producer to Sequential consumer No producer or no consumer (typically. transmitted. as illustrated in the example on the right: The degree of parallelism is identified in brackets after the operator name. or translated into any language in any form by any means without the written permission of IBM. The memory used by deadlock-prevention BufferOps can be calculated based on the number of inserted BufferOps. 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. and the next “ds1”. . Producers and consumers may be either persistent (on disk) Data Sets or parallel operators. The partition type is associated with the first term. Data Sets are identified in the first section of the parallel job score. in the example on the right.

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

stored in a retrieval system.Information Integration Solutions Center of Excellence the entire structure needs to be loaded before actual lookup processing can begin.ds) on nodes ( node1[op1. 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.ds)} . you can also determine the delay associated with loading the lookup structure.p0] )} op1[1p] {(parallel delete data files in delete temp.ds) on nodes ( node1[op2. transmitted. In a similar way.p0] )} op2[1p] {(sequential delete descriptor file in delete temp.Delete Data Files ->eCollectAny op2[1p] (sequential delete descriptor file in delete temp.ds} It has 3 operators: op0[1p] {(sequential Row_Generator_0) on nodes ( node1[op0. 2006 152 of 179 © 2006 IBM Information Integration Solutions.p0] )} Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.ds) . 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. All rights reserved. .Delete Descriptor File ds1: {op0[1p] (sequential Row_Generator_0) -> temp.

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: .Information Integration Solutions Center of Excellence Appendix D: Estimating the Size of a Parallel Data Set For the advanced user. No part of this publication may be reproduced. 2006 153 of 179 © 2006 IBM Information Integration Solutions. All rights reserved. or translated into any language in any form by any means without the written permission of IBM. As can APT_Transfer::getTransferBufferSize(). stored in a retrieval system.add (# nullable fields)/8 for null indicators . if you have a transfer that transfers all fields from input to output. . 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. 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. transmitted. Parallel Framework Red Book: Data Flow Job Design July 17. transcribed.5 bytes per field) Using the internal DataStage Enterprise Edition C++ libraries.

or to change the default behavior of specific DataStage Enterprise Edition stages. 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. By default. See section 4.2: Default and Explicit Type Conversions 2. to assist in debugging.Information Integration Solutions Center of Excellence Appendix E: Environment Variable Reference This Appendix summarizes the environment variables mentioned throughout this document. These variables can be used on an as-needed basis to tune the performance of a particular job flow. 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. 1. or pads.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. transcribed. 2006 154 of 179 $APT_IMPORT_REJECT_STRING_FIELD_OVERRUN S 1 (DataStage v7. but there is a small performance penalty from increased I/O. Setting this environment variable directs DataStage to reject Sequential File records with strings longer than their declared maximum column length. All rights reserved. a variable-length string field to a fixed length (or a fixed-length to a longer fixed-length). Defines size of I/O buffer for Sequential File reads (imports) and writes (exports) respectively. NOTE: The environment variable settings in this Appendix are only examples. imported string fields that exceed their maximum declared length are truncated. In some disk array configurations. 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. An extensive list of environment variables is documented in the DataStage Parallel Job Advanced Developer’s Guide. Set values that are optimal to your environment.1. Setting this value to a low number (such as 1) is useful for realtime applications. Increasing these values on heavily-loaded file servers may improve performance. July 17. setting this variable to a value equal to the read / write size in bytes can improve performance of Sequential File import/export operations. transmitted. . stored in a retrieval system. Default is 128 (128K). with a minimum of 8.

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

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

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.

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

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

transmitted. 6 There is an exception to this rule: If your hash key has only one value. No part of this publication may be reproduced. 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. Minimizing Record Movement for Maximizing Performance Now we have covered the basic rules and mechanics for hash-partitioning and sorting. . transcribed.Information Integration Solutions Center of Excellence 3 102 Example: 6 rows in 2 partitions. 2006 162 of 179 Parallel Framework Red Book: Data Flow Job Design © 2006 IBM Information Integration Solutions. let’s look at how we can capitalize on these behaviors for performance benefits. 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. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. Here is another possible outcome: Also: Consider the result of running the same job with the same data. stored in a retrieval system. July 17. but a different number of partitions.

No part of this publication may be reproduced. and append these values to the original data. transmitted. All rights reserved. as well as the average transaction amount per-item. or translated into any language in any form by any means without the written permission of IBM. 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. 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.Information Integration Solutions Center of Excellence Scenario Description: Our customer is a national retail business with several hundred outlets nation-wide. thereby increasing the value of this exercise. . 2006 163 of 179 © 2006 IBM Information Integration Solutions. 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. 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. transcribed. per store for all stores in the nation. They wish to determine the weighted average transaction amount peritem nation-wide.

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

5 50. .5 39.5 310. not Hash. It does this by keeping running totals in memory for the aggregation for each output group.25 26.5 3. transcribed.5 23 23 23 23 26. transmitted. it consumes an amount of RAM proportionate to the number of output rows and the number of columns involved in the aggregation. or translated into any language in any form by any means without the written permission of IBM.5 16. however.25 Store Average Item Transaction Amt 2 2 5 5 6 6 July 17. Partition 0: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35. JoinSourceToAggregator_2 produces the final result: the original input Data Set with two columns appended (‘Average Item Transaction Amt’. 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.5 35. The hash method only requires the input data to be hashed. 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.5 310.25 16.5 35.5 180.5 3.5 3.25 180.Information Integration Solutions Center of Excellence NOTE: For this scenario.5 180. in return.5 39.5 180.5 46 46 6. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3. stored in a retrieval system. No part of this publication may be reproduced.25 26.5 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. you will need to set the Aggregator’s “Method” to Sort. 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. All rights reserved. Therefore.5 50.5 6. it does not guarantee output order. The sort method requires the input data to be hashed and sorted.25 26.5 6.5 35.5 Store Average Item Transaction Amt 28 28 43 43 6.

5 33.5 61. or translated into any language in any form by any means without the written permission of IBM. transmitted.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. Here’s a screen shot of this. 3. A deeper examination (of the score dump. and 6.25 38.75 26. i. 3. No part of this publication may be reproduced.25 44.25 16.e. and 3. appended to the end of this document for masochistsiv) might suggest: 4.75 38.25 44. .: record adjacency assumes partition concurrency8. An examination of the job above would suggest: 6.5 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. All rights reserved.75 38. more efficientv.25 44. transcribed. 2006 166 of 179 © 2006 IBM Information Integration Solutions. Parallel Framework Red Book: Data Flow Job Design July 17.25 44.5 27 27 61. and 1. A much better answer is: 1. even sequential job execution does not alter the answer for the sort-related questions.75 38. This is certainly an improvement on the previous answer.5 33. solution (score dump also attached below vi): 8 Records cannot be adjacent if they are not in the same partition. as only partition concurrency is affected by sequential execution. stored in a retrieval system. 6.

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

instead of buffering the entire Data Set. However since we sub-sorted the data before this aggregator (unlike NationalAverageItemTransactionAmt). . All rights reserved. ALL extant values of the remaining columns are already in the same partition. however dreadful). The output of StoreAverageItemTransactionAmt contains the other column we need to append to our source rows. we will have to prep the output from the first join to account for Parallel Framework Red Book: Data Flow Job Design July 17. This means that the output of the DistributeCopiesOfSourceData and NationalAverageItemTransactionAmt are already hashed and sorted on the keys needed to perform JoinSourceToAggregator_1. even for ‘fast’ disks). 2006 168 of 179 © 2006 IBM Information Integration Solutions. If the group size was only several hundred records. especially in MPP environments where repartitioning implies network I/O). The previous two paragraphs contain two key concepts in Enterprise Edition (pun fully intended. which is what makes this sort possible w/o re-partitioning (which is also quite expensive. but the entire Data Set was 100 million records. we have instructed the sort stage that the data is already sorted on ItemID and TransactionDate (as always with sorting records. 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. This accomplishes the first goal. transcribed. but the group keys will force the proper order). 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). No part of this publication may be reproduced. transmitted.Information Integration Solutions Center of Excellence As you can see. 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). Also worth noting here: because we already hashed the data on ItemID and TransactionDate. granted. Getting back to the aggregators. to append a column representing the national (weighted) average item transaction amount. key order is very important). 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. the rows will come out in the same order they went in (different rows. stored in a retrieval system.

Information Integration Solutions Center of Excellence the new row ordering of StoreAverageItemTransactionAmt. useful for preserving previous sort orderings). transmitted. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.25 44.75 38.5 16.5 33.25 44.75 38.5 33.5 3. stored in a retrieval system. Output from above solution: Data Set 3: PeekFinalOutput. All rights reserved.25 3.25 16. 2006 169 of 179 © 2006 IBM Information Integration Solutions.25 38.5 61.25 16.75 38.25 44.75 Store Average Item Transaction Amt 2 2 6 6 5 5 26.5 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. Partition 2: 16 Rows Parallel Framework Red Book: Data Flow Job Design July 17.5 3. or translated into any language in any form by any means without the written permission of IBM. and it is enabled by default. 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. .25 44.5 26. No part of this publication may be reproduced.5 27 27 61. as it is much more expensive than non-stable sorts. transcribed.5 16.

000. a single partitioner. That is a LOT of saved processing power. but there is a critical difference.5 23 23 180. Imagine a job with 100 million records as the input. and only hashes.25 180. The second solution only sorts (on disk) 100. stored in a retrieval system.5 Store Average Item Transaction 28 28 43 43 6.5 39.5 6. a savings of 400.25 26.5 35.5 6. No part of this publication may be reproduced.000 additional record movements—half of them involving disk I/O—for a 100 million record input volume. With the initial solution.5 23 23 26. 2006 170 of 179 © 2006 IBM Information Integration Solutions.000 records. and two sub-sorts—a much more efficient solution for large data volumes.25 26. Parallel Framework Red Book: Data Flow Job Design July 17.000 records in addition to hashing 300.5 35. .000 records.5 50.5 180. 100.5 39. It looks very similar to the first solution.000. 000.5 This solution produces the same result but is achieved with only one complete sort. transmitted.5 310.5 50.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. transcribed.000 records.5 180. we had to sort (on disk) 300.25 26. There is an even more efficient solution.5 35.000. All rights reserved.5 310.5 46 46 6. or translated into any language in any form by any means without the written permission of IBM.000.

This is functionally equivalent to doing a sub-sort right before the StoreAverageItemTransactionAmt aggregator. However it will not create additional processes to handle the records and re-order them.Information Integration Solutions Center of Excellence NOTE: In this job. transcribed. No part of this publication may be reproduced.25 Store Average Item Transaction Amt 2 2 5 5 6 6 26.5 16.25 16.5 26. Partition 1: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 3.a significant savings.25 44. The difference is on DistributeCopiesOfSourceData: Here. This is a potentially huge savings on large data volumes (remember the previous example). we saved a sub-sort on 100 million records . and automatic partition insertion must be turned off vii. or translated into any language in any form by any means without the written permission of IBM. All rights reserved. but NOT to use it for hashing. Comparing the efficiency of this solution with that of number two. Here is the output from this version of the job.5 3.25 16.25 44. we have chosen to use the StoreLocation column as a part of our sorting key. Looks a lot like solution 1. stored in a retrieval system. for the same reasons.5 3. Also. transmitted. the need for the second sort on the output of the JoinSourceToAggregator_1 is not needed.25 16. Data Set 4: PeekFinalOutput. automatic sort insertion.5 3. . 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. except w/o the sort on the output of JoinSourceToAggregator.5 27 27 July 17.

2006 172 of 179 © 2006 IBM Information Integration Solutions. sort. No part of this publication may be reproduced.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.25 38. stored in a retrieval system. 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 46 46 6.5 35. transmitted.25 44. or translated into any language in any form by any means without the written permission of IBM.5 180.5 23 23 23 23 26.25 26.5 35.5 180.5 33. transcribed. in addition to the heavy penalty paid in disk I/O for using a full sort.75 38.75 38.75 61.5 Store Average Item Transaction Amt 28 28 43 43 6.25 26.25 180.5 35.5 33. All rights reserved. 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.5 6.5 39.5 50. Partition 2: 16 Rows Store Location Item ID Transaction Date Transaction Amt National Average Item Transaction Amt 35.5 310.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 50.5 44 44 PeekFinalOutput.25 26.75 38.5 310. .5 39. by definition.5 180.

the link sort in solution 2. transcribed. 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. No part of this publication may be reproduced. All rights reserved. This allows down-stream stages to be processing data during the sorting process. no rows have left yet. 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). we are sub-sorting the data we sorted in the previous diagram). instead of waiting until all 40 million records have been sorted (in this instance. requiring a large amount of scratch diskx. . Parallel Framework Red Book: Data Flow Job Design July 17. This is because a standard sort requires all rows to be present in order to release the first row.Information Integration Solutions Center of Excellence As you can see. This situation is analogous to all of the sorts in solution 1. stored in a retrieval system. transmitted. 2006 173 of 179 © 2006 IBM Information Integration Solutions. and the link sort in solution three. although ~5 million records have entered the sort.

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

p2] )} op2[3p] {(parallel APT_HashedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op2.[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] )} op1[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op1.p2] )} op5[3p] {(parallel APT_TSortOperator(1)) on nodes ( node1[op5.p1] node1[op1.p2] )} op4[3p] {(parallel APT_TSortOperator(0)) on nodes ( node1[op4.p1] node3[op5.p0] node2[op2.p0] node2[op3.p0] node1[op0.DSLink18_Sort)} ds12: {op10[3p] (parallel JoinSourceToAggregator_2. key={ value=StoreLocation.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.p0] node1[op1. key={ value=TransactionDate }.p2] )} op6[3p] {(parallel buffer(0)) .DSLink2) on nodes ( node1[op0.p1] node3[op4. subArgs={ cs } }.p2] )} op3[3p] {(parallel APT_HashedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op3.p0] node2[op5. subArgs={ cs } } })#>eCollectAny op10[3p] (parallel JoinSourceToAggregator_2.p1] node1[op0.p1] node3[op2.p0] node2[op4.p1] node3[op3.

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

p2] )} op4[3p] {(parallel APT_SortedGroup2Operator in NationalAverageItemTransactionAmt) on nodes ( node1[op4.p2] )} op3[3p] {(parallel SubSortOnStoreLocation) on nodes ( node1[op3.p2] )} op1[3p] {(parallel DistributeCopiesOfSourceDta.p1] node1[op0.p1] node3[op4.DSLink2) on nodes ( node1[op0.p2] )} op5[3p] {(parallel APT_SortedGroup2Operator in StoreAverageItemTransactionAmt) on nodes ( node1[op5.[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.p0] node2[op2.p0] node1[op0.p1] node3[op3.p0] node2[op3.p0] node2[op4.p0] node2[op5.p1] node3[op1.DSLink2_Sort) on nodes ( node1[op1.p0] node2[op1.p2] )} op2[3p] {(parallel DistributeCopiesOfSourceDta) on nodes ( node1[op2.p1] node3[op2.p2] .p1] node3[op5.

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

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

Sign up to vote on this title
UsefulNot useful