Document toolboxDocument toolbox

Sort Stage

Summary

The Sort stage sorts data according to the provided specifications.

IBM Documentation

Server Stage: Sort Stages - IBM Documentation

Parallel Stage: Sort stage - IBM Documentation

Conversion Notes

Structural changes

A Sort stage supports a configuration of one input link and one output link, in both the Server and Parallel versions. As such, there are no scenarios that call for a structural change to jobs due to the presence of a Sort stage.

Memory Usage

In Server edition, Max Rows In Virtual Memory denotes the number of records to sort in-memory before writing to temporary files, and defaults to 10K records. The Parallel Sort stage instead uses the property Restrict Memory Usage to set a memory threshold as a number of Megabytes, defaulting to 20 MB. A suggested maximum limit for this value is set at 500 MB for Windows installations. Due to there being no logical conversion between these two values (records and Megabytes), and the differences in memory usage between the Server and Parallel engines, the Restrict Memory Usage property will not be set by S2PX and so the generated parallel Sort stage will operate with the default setting of 20 MB.

Sort Keys

The Server version of the Sort stage lists the sort keys as a string in the Sort Specification property, documented here. For each key in the Sort Specification a appropriate Key value is generated in the Parallel Sort stage properties.

Server features not supported

Feature

Asset Query (?)

Comment

Feature

Asset Query (?)

Comment

File Management options

Sort with Temporary Directory

Not supported.

The Parallel Sort stage writes the temporary sort files in a location pre-determined.

The related Server properties have no equivalent in Parallel are:

  • Temporary Directory

  • Max Open Files

External Map Files in Sort Specifications

Sort Specification with External Map File

Not supported.

Here is an example of a Server stage Sort Specification with an external map file:

REGION asc, CUSTOMER ASC C:\USER\CSM, SALE_PRICE DSC

This map file (C:\USER\CSM) allows custom sort ordering to be mandated, and would be useful for sets of codes that are not ordered alphabetically. This feature is not supported within the Parallel Sort stage design.

The related Server properties are not relevant in Parallel:

  • Column Separator

  • Escape Character

Tracing Level granularity

n/a

Tracing Level in the Server stage is an integer that behaves like 4 bit flags. There is no parallel equivalent. There is no asset query for this property as its lack of support in Parallel Jobs has no functional implications.

 

 

 

© 2015-2024 Data Migrators Pty Ltd.