User inserted sort

Post questions here relative to DataStage Enterprise/PX Edition for such areas as Parallel job design, Parallel datasets, BuildOps, Wrappers, etc.

Moderators: chulett, rschirm, roy

Post Reply
boxtoby
Premium Member
Premium Member
Posts: 138
Joined: Mon Mar 13, 2006 5:11 pm
Location: UK

User inserted sort

Post by boxtoby »

Hello,

I have a job which is identical in development and production.

In production it works because the partitioning behaves correctly on the two join stages in the job.

In development the job doesn't behave correctly and produces slightly different results each time I run the job.

I have actually fixed the problem in development by adding explicit sorts and partitioning which is not entirely unexpected, but I would like to know wht the two environments are behaving differently.

In fixing the job in development, in addition to the sort and partitioning that I added, I included the APT_NO_PART_INSERT and APT_NO_SORT_INSERT environment variables set to true.

The same set of changes worked in production, not surprisingly.

Can anyone shed any light on why development and production should behave differently? The above two environment variables defaulted to true when inserted in to the job so I don't think it's that. Is there another environment variable I could check, perhaps?

Many thanks and look forward to hearing from you.

Kind regards,
Bob.
Bob Oxtoby
qt_ky
Premium Member
Premium Member
Posts: 2895
Joined: Wed Aug 03, 2011 6:16 am
Location: USA

Post by qt_ky »

Check that your C++ compiler versions, patch levels/PTF levels are identical across environments, assuming you have a compiler in production. Some places do not.

If your development compiler version got changed for any reason (it becomes an unsupported compiler version at that point), then very strange things can start happening, including unexpected and incorrect results.
Choose a job you love, and you will never have to work a day in your life. - Confucius
Post Reply