DSXchange

Can't you just sort the result again, by emplid then by rank in descending order? Or run it through an Aggregator stage grouping by emplid, name and groupid and returning max(rank).

Incidentally, choose a different term. What you have is not a rank, it's merely a counter.

Why not enable server side tracing and do an export from Manager, then disable server side tracing and inspect the trace file so you can see how DataStage does it? It might give a clue. Then again, it might not.

Two possibilities. 1. Your Repository indexing is out of date. Repair with DS.REINDEX when you have exclusive access to the project. 2. The job has already been deleted. Verify with DS.CHECKER when you have exclusive access to the project. If you want me to answer your questions immediately when you...

Reject the requirement. It can not be reliably achieved.

Learn English. The second person personal pronoun is "you", not "u". Plz, I believe, is a city in Borogravia.

If you have Enterprise Edition the parallel job Sequential File stage has the capability you ask for (a file name column). Otherwise you will need to construct a loop in a job sequence to process the list of file names, and pass each individual file name to the job, which can then create a file name...

Learn patience. ProfileStage can not let you view the sample until it has finished generating the sample.

Continuously-running jobs are, in general, a bad idea except, perhaps, when exposing your job as a web service via RTI stages (SOA edition). Let the jobs finish at least once per day so resources can be cleaned up. Also provide a mechanism so that you can instruct the job to finish on the next cycle...

It's "sequence", not "sequencer". Job exit status is 0 = running, 1 = finished OK, 2 = finished with warnings, etc. You need to create a custom trigger on the Execute Command activity so that it does not assume that a non-zero exit status is automatically a failure.

... or use lookup_string_from_int16() function in the Modify stage.

Neither is "better". Which is best suited to your requirements? Is the target database available whenever you need it, or should you store the data so they can be loaded when the database is available? In that case, a Data Set is probably preferable to a file, since it preserves the parallel nature ...

DataStage without NLS enabled requires what you call "American settings".
DataStage with NLS enabled can handle any character known to modern business data processing.

All warnings are displayed in the job log, unless you have a maximum number of warnings set when the job is submitted to be run (usually 50). This can be set in the Job Run Options dialog when you submit a job to be run from Director, or using the -warn option in the dsjob command.

Can you please check the STRIPLIST - that's the only explanation I can think of for that behaviour, unless the sort itself ignores "punctuation" characters, which I think would be unlikely. Have you involved your official support provider?

Everything in C starts from 0. Get used to it.

Ask your DB2 DBA about the bind20 utility.