Is the Output of Sort/Join dependent on any system config?

Sanjith · New User Joined: 14 Jul 2021 Posts: 1 Location: UK

Dear All,

I have been observing different outputs for the same Sort/Join Step across 2 separate Jobs. The input files/sort card are all exactly same. The only difference is that one process is running outside of the Production environment and the other is locally on my spool under my ID.
I can't find any logical reason for such a difference in output.

Skeleton structure of my Step

sergeyken · Senior Member Joined: 29 Apr 2008 Posts: 1751

Neither Region, nor Memory Space.
Nor DD statements, too.

Only - the code itself (SORT statements, not shown!), and the input data too (not shown as well!).

Also need to analyze carefully all the messages from SORT utility, in both cases.

In theory, a SORT parameter STOPAFT=700 can be introduced in one of your systems; but this probability is very low. You can also find it from the detailed SORT messages log!

Rohit Umarjikar · Posted: Wed Jul 14, 2021 7:50 pm

I am sure only variable between two is input Dataset. I would suggest to pick up the same Generation or PS used outside production and then see if they are really 100% match.
Second, I don't think DFSORT version difference would impact this but no harm to compare that as well.
Third, remeber you coded DISP=OLD, try creating NEW Data set and compare at both places or look for the SORT messages for IN and OUT counts in the job log.