Bulk Data Movement Guide

10.5
- 10.5.6
- 10.5.5
- 10.5.4
- 10.5.3
- 10.5.2
- 10.5.1
- 10.4.1
- 10.4.0 HotFix 1
- 10.4.0

Back Next

Rules and Guidelines for Multiple-Record Writes to IMS Unload Data Sets

Review the rules and guidelines that pertain to multiple-record writes to IMS unload data set targets before you configure a workflow that uses this feature.

For more information, see the

PowerExchange Interfaces for PowerCenter

Source and Target Files

The following rules and guidelines apply to sources and targets:

If the target is an IMS unload file, the source must also be an IMS unload file.

The source IMS unload file can be in any format that PowerExchange supports as a data source. For more information, see Using IMS Unload Data Sets as Data Sources.

Only standard IBM unload formats are supported for targets. BMC or partitioned unload files are not supported.

The unload file is written in Fast Path database format if the data map specifies a Fast Path DEDB database. Otherwise, the file is written in the standard IBM unload format.

If you run the IBM unload utility to unload data from multiple partitions, the partitions are interleaved in the output unload file. Before you use the unload file as a data source in a workflow that performs a multiple-record write, you must sort the unload file to sequence the partitions. Alternatively, you can add the following control card to the unload job:

TASKCTL=(1)

If you unload segments for multiple areas of a Fast Path DEDB to a single unload file, you must sort the unload file before you can use it as a source in a workflow that performs a multiple-record write.

Data Maps, Source Definitions, and Target Definitions

The following rules and guidelines apply to data maps, source definitions, and target definitions:

You must create new source and target definitions for multiple-record writes. Existing definitions without sequencing information do not work for multiple-record writes. You can continue to use existing definitions for workflows that do not use multiple-record writes.

The source and target definitions for IMS unload files must be based on the same data map.

The data map must define the SSID and DBD name. These fields might be blank on existing DL/1 or ODBA data maps. If you try to import an IMS data map with a blank SSID or DBD field into a source definition for which you enable multiple-record writes, PowerExchange issues messages PWX-03611 or PWX-03612.

Verify that all tables that form a complete hierarchy are included in the data map. The import process does not verify hierarchical completeness.

To ensure that all of the required tables in the data map are included in the source definition, verify that they are defined in the data map as simple tables. A simple table
is based on a single record.

If the data map includes complex tables, PowerExchange does not include them in the source or target definition. A complex table
is based on more than one record.

You cannot select which tables in the data map to import. PowerCenter imports all of the simple tables in the data map.

PowerExchange excludes any user-defined, RBA-related, or CCK-related fields that are in the imported IMS data map from the source definition.

The data map must have exactly one table for each record.

Mappings and Transformations

The following rules and guidelines apply to PowerCenter mappings and transformations:

Mapping transformations might drop rows. In this case, ensure that the mapping does not create any orphan rows. When a multiple-record write workflow runs, it drops orphan rows instead of writing them to the target.

The mapping must not alter the position or the name of the generated sequence fields or drop them.

Transformations that insert records into a mapping are not supported.

Transformations that might alter row IDs, including the Sorter, Joiner, and Rank transformations, are not supported.

Connections

The following rules and guidelines apply to connections and connection properties:

You can use multiple-record writes with PWXPC NRDB Batch connections only. PowerExchange ODBC connections are not supported.

Select

Filter After

for the

Offload Processing

source and target connection attributes. PowerExchange and PWXPC perform offload processing on the Integration Service machine before writing the output records to

z/OS

If you select a different value, PowerExchange changes the value to

Filter After

However, if RBAs are being read using the GetDatabaseKey or GetIMSRBAByLevel function from an IMS unload file in the BMC LONG format, you must select

for the

Offload Processing

source connection attributes. Otherwise, the session will fail with message PWX-03803.

Multithreaded processing is supported for IMS source connections if you set the

Offload Processing

connection attribute to

Filter After

and set the

Worker Threads

connection attribute to a non-zero value.

Multithreaded processing is not supported for IMS target connections. If you set the

Worker Threads

connection attribute, the setting is ignored.

Select

Off

for the

Confirm Write

source and target connection attributes. If you select a different value, PowerExchange changes the value to

Off

In certain cases, you might need to change the value of the

CSQ_MEMSIZE

parameter of the

PWX Override

connection attribute. This parameter defines the maximum amount of memory that the cache can consume for multiple-record writes.

Session Properties

The following rules and guidelines apply to PowerCenter session properties:

Pipeline partitioning is not supported for the reader or writer if you defined sources or targets with sequencing enabled.

You must specify the

IMS Unload File Name

property for IMS unload sources and targets.

For sources, the

Flush After N Blocks

property specifies the maximum number of block flushes that can occur across all groups without a given block being flushed. Define this property to ensure that all blocks are flushed at regular intervals.

Rename Saved Search

Table of Contents

Bulk Data Movement Guide

Bulk Data Movement Guide

Rules and Guidelines for Multiple-Record Writes to IMS Unload Data Sets

Rules and Guidelines for Multiple-Record Writes to IMS Unload Data Sets

Source and Target Files

Data Maps, Source Definitions, and Target Definitions

Mappings and Transformations

Connections

Session Properties