DSXchange

I believe that the right place to introduce "correct" practices for this site is a Terms and Conditions screen, much like one gets on banking, airline, insurance, etc., sites. People signing up (that is, joining) would have to check a box indicating that they agree to the terms and conditions. These...

Right below the Generate button in Designer is another button called Upload. This is how you get the COBOL and JCL to the mainframe. After that, it's someone else's problem; DataStage does not - can not - control execution on the mainframe. And the politics of getting permission to compile and run o...

It might be worth validating the jobs in the new environment, to make sure that parameter values are correctly set. If they aren't, go to Director and choose Set Defaults from the Job menu.

In parallel jobs you can use a Lookup stage against a Lookup File Set stage or performing a sparse lookup against a database stage (e.g. DB2 or Oracle). In either case the lookup key column is mapped automatically with no requirement for manual intervention. No such capability exists in server jobs ...

LOGTO UV is deliberately blocked in the Administrator client Command window (for reasons I can't fathom), but can be used if you're in a dssh environment on the server.

Welcome aboard. :D Your choice of subject line won't help anyone else searching for assistance with this particular problem; please try to be more meaningful with future Subject lines. You should be able to identify the rows with three cascaded pairwise joins (Join stages). File1(Key1) = File2(Key3)...

Umm, "abc" doesn't really have two decimal places?!!

Maybe the data type requires precision AND scale, for example decimal[10,2].

That's how I'd be thinking too. Just as in Oracle, your date must match your date picture (which is yyyy-mm-dd). So try 1900-01-01. If that fails try "1900-01-01".

... and I managed to sleep through the whole drama! :D

Well done on solving it! We'll get you weaned onto data types yet!

Check to see whether a file has been written into the &PH& directory. Or try resetting the job in Director, to see whether a "from previous run..." event picks up the output from your printf() calls. Prefer SIGTERM to SIGKILL at all times. It allows a grace time for buffered I/O to complete ...

For removing everything between parentheses, and assuming there is only one set of parentheses, you can use an expression. If there is more than one set of parentheses, you need a routine that iteratively processes the string using the same logic for each set. Extract everything up to and including ...

Is Oracle busy doing other stuff? The DBA can help you to track down what's being asked of Oracle at any particular time, and maybe why timeouts are occurring.

D'you think we (oops, I mean the dudes who fund this site) should sell such success stories to IBM or whomever?

In server jobs (yes, I am aware Brad uses parallel jobs) there is a technique using a "range conversion" with Oconv. This returns the original string if it falls within one of the ranges in the second argument, or an empty string otherwise. To accomplise Brad's example you would use: Oconv(Argum...

You haven't given us enough in the two excerpts. We need to see the preceding IN= and OUT= messages (and everything between them) to understand what DataStage was attempting to do that took so long.