DSXchange

Is it possible for server resources to impact a join without throwing an error? I have a job that left outer's on two datasets that produces incorrect results every now and then. The join matches on key's that do not exist and thus I'm getting incorrect values from the right columns.

I've verified that the partitioning is correct. Subsequent job runs produce the desired result. Has anyone experienced a similar issue?

I don't believe it's possible.

Create a reproducible case then report the alleged bug to IBM through your support provider.