DSXchange

Change your output to a dataset. Is the behaviour the same? If not, then the cause is on the database side.

I had a similar issue under 7.5.2 and AIX but in my case neither of the two setting seemed to make a difference and the job was doing a commit at the end, as if the frequency was set to 0. We had opened up a support call and after we had changed the job around (for other reasons) it suddenly started...

That particular post only had a couple more words to it, but it is certainly worthwhile getting a premium membership.

Did you create your range map with the write-range-map stage?

Nope, nobody can.

It is a user-written function in your project, so you should look at the job code and figure out what it does.

How can we do that since we are awaiting a description of what you are doing from yourself.

This doesn't make sense. Can you post the exact compile-time error message, copied over with cut-and-paste?

p.s. indiscriminetely killing processes as suggested by Madhav_M is a very bad idea and can, in the worst case, lead to you have to restart the DataStage Engine!

Pk39565 - I can guess that you did what everyone here recommends that you not do - you used a "kill -9" instead of a normal "kill". You now have a lock without an associated process. Once again, this is a case where I can only recommend that you have the deadlock daemon running; in which case you co...

What size did you set it to? 2147483647? How long is your maximum string length actually going to be - setting it to 2Gb could be overkill.

The "EREPLACE" function doesn't exist in the PX transform stage, so the external sed function, perhaps in a wrapper, will work but might be just as slow as using a BASIC transform stage. In that case, it would be better to evaluate writing this job in Server.

Sometimes that cannot be done due to other constraints.

man pages for the iconv function are present on AIX as well.

Yes, you need to set NLS_LANG to reflect the Oracle database language.

With a normal lookup the lookup stage will have a link from a column in the top left to a column marked as "key" in the bottom left. Are both those columns called "Id" in your job? What happens if you change the column name of at least one of them as suggested above?

I can't check DS right now, but think that there was a debugging type switch in the stage to give you more error output.