DSXchange

The file will always be created.

Guessing "black" = "blank".

Sequence job. Use a routine to check for an empty file and skip the FTP job if needed. Notification for empty file or job failure.

That would be a "fixed width" file. In other words, there are no field delimiters and any reader program has to know the length of each field to know where any given one starts and stops.

Simplest way to find out is to try it... what happens? $CommandOutput will capture anything the command writes to standard out, i.e. outputs to the 'screen'.

You open the file with what? And when you 'save it back' what has been changed about the file? Do you have the same issue if you use split rather than csplit?

How would you do this outside of DataStage? In Oracle, for example, one would create a dblink between them, there must be a similar technique for Teradata.

There isn't really a 'reset issue' here as it's not hurting anything. Suggest you ignore that and concentrate on solving the duplicate rows problem so the job doesn't abort. If it really bothers you, work on properly setting up DataStage for DB2. It sounds like what you did by setting those environm...

It's great that you want to come here to help and all, but the link to a porn site isn't really appreciated nor appropriate.

Only in this forum.

viewtopic.php?t=124402

Sorry, for what you've added you're going to have wait for Grognard help.

I'm curious why all the shenanigans, why not just stick with the job(s) you've obviously already built and which are apparently 20x faster?

Just you.

Doesn't really surprise me, seeing as how you are basically printing a 'report' to disk. Those numbers in the SETPTR function are line length, page length, top margin, bottom margin, etc. The UniVerse User Reference has all of the gory details.

Me, I'd just stick to using a job.

Missed it by *that* much.

So much for keeping the use of code tags in mind. Go back, edit the post and add them, please. For grins if nothing else.

Yup, edited my previous post to include the "why".