DSXchange

I meant what I said - leverage a sequence object, say in a database or job. A Sequence job could control that, sure, and pass it to the job that needs it.

If your 'running number' needs to be incremented each time the job runs, store it in a sequence of some kind. Increment it when you build your parameter value. Heck, even a flat file would work.

Hence my first question.

Ah... the script resets any non-success job immediately so all the script can report is success - either the job finished successfully or the reset did. That's backwards in my opinion, you need to look at the failed job before resetting it. The script can check before starting the job and reset befo...

That's still the prevailing wisdom. The O/S will 'reap' dead processes when it is possible, however any that you see in that undead state still require a reboot to clear.

Take it at the annual IOD... err... Insight conference in Las Vegas. From what I recall you can retake for free while there.

What O/S? Google "zombie process" to get an idea what's going on and how to handle them. There's nothing automatic about it, I'm afraid.

Not sure what you mean by 'in DataStage'. We've designed our Type2 tables to include metadata columns that let us know what was inserted / updated per load but that allows collection of the information after the fact by querying the targets. We also use separate target stages for 'insert inserts' an...

Never mind the whole 'relative path' thing, while fine advice that's obviously not the issue here. Your error is an odd one and generally indicates a compatibility issue - say you bring over the command from another version of UNIX or one compiled with the 'wrong' compiler for the current O/S you co...

<post moved to correct forum>

Post your script... what's on line 9? Looks like you're using a relative path, make sure your script cd's to the correct place first or switch to a full path.

Those are the 'correct' counts, technically. However, perhaps you could subtract the updates from the inserts to get the insert count for new records only in this case.

If you search, there are 22 other posts with that same 'no rows found' DB2 message in them. Perhaps this one helps? viewtopic.php?t=124265

Going to sticky this topic for a little while so it doesn't roll off... it can start its slide into the history books in March.

Someone wrote the DS_RunJob.sh command for you, I would imagine. Don't recall any such thing shipping with the product, unless it is what the Director uses under the covers in crontab? Open it up and have a look at it, I'll wager it uses the dsjob command line function. Could be pretty bare bones or...

Amin wrote:We need to implement this using Datastage job (Parallel or Sequence) and that Job is place before and after the orgional job in sequence

Really? You've just tripled the number of jobs you run daily. Actually quadrupled.