APT_INSERTION_SORT

san_dwh · Post by **san_dwh** » Fri Jan 21, 2011 12:07 am

What is the exact purpose of APT_INSERTION_SORT

ray.wurlod · Post by **ray.wurlod** » Fri Jan 21, 2011 12:32 am

There is no environment variable of this name in DataStage.

Is this an interview question?

XRAY · Post by **XRAY** » Fri Jan 21, 2011 1:21 am

Do you mean APT_NO_SORT_INSERTION ?

san_dwh · Post by **san_dwh** » Fri Jan 21, 2011 3:18 am

To ray
yes

san_dwh · Post by **san_dwh** » Fri Jan 21, 2011 3:23 am

@XRAY.

ya i think. I Don't about this. will you please explain something

XRAY · Post by **XRAY** » Fri Jan 21, 2011 6:24 am

Having set APT_NO_SORT_INSERTION = FALSE , datastage will insert sort to your job though you have not explicitly defined it in your job.

There is an earlier discussion on this topic

http://66.34.129.47/viewtopic.php?p=386 ... 365c03d1fd

"Join stage

By default, the Join stage sorts data on both input links. This sorting results in output data that is sorted on the join keys. If the InfoSphere DataStage project environment variable APT_NO_SORT_INSERTION is set, the sort is suppressed. Pushing a Join stage into a database source always generates an ORDER BY clause for the sort keys, as if APT_NO_SORT_INSERTION was not set. "

jwiles · Post by **jwiles** » Fri Jan 21, 2011 9:29 am

As was stated in that thread, the quoted documentation that was referenced by the OP was specifically for InfoSphere Balanced Optimization, an optional licensed add-on to Information Server. The context in which APT_NO_SORT_INSERTION is mentioned here can be misleading as to it's actual purpose.

san_dwh,

APT_NO_SORT_INSERTION, when set to 1, prevents the automatic insertion of sort operators into a parallel job at runtime. This is documented in the Parallel Job Advance Developer's Guide included with Information Server, as well as in the online Information Center documentation for Information Server found on the IBM website and in numerous threads here on DSXchange.com

Regards,

san_dwh · Post by **san_dwh** » Mon Jan 24, 2011 4:32 am

thank you

ray.wurlod · Post by **ray.wurlod** » Mon Jan 24, 2011 2:05 pm

In an interview I'd still go with "there's no environment variable of this name". It's good to see the PHB discomfited.