DSXchange

If your derivation ends with then the 20 will be added to the end of the string.

This type of error message can take a while to track down, particularly in bigger jobs with local and/or shared containers. Sometimes using "$APT_PRINT_SCHEMAS" can help narrow down the location, since you can track which links between stages contain "CURRENCY" and which don't.

anbu has stated it correctly, in order to get a UNION of two streams you use a FUNNEL stage (I mistakenly type "join"), then to be compatible with the SQL union functionality you need to remove any duplicates.

I'll assume that this is not an interview question and suggest that you look at the Parallel Job Developer Guide chapter for the join stage.

Remove all stages except your dataset -> copy -> peek stages. Is the error still present? If yes, add "$APT_DUMP_SCORE" and set to "true" and run your job. Examine the score to see if that helps you, otherwise post it here.

Using a server job or, better yet, logic in a job sequence is going to be much simpler than attempting to do in PX with a wrapped stage. Is your EMP_ID file a sequential one with just the EMP_IDs in it and no other information? If so, you could read the file into a variable (just "cat" the...

You will need to use calls to DSGetJobInfo(), DSGetStageInfo() and DSGetLinkInfo() to get this information. Are you sure that you want to program this yourself or can use existing code such as that from Kim Duke?

Not too simple to achieve, but there are ways - particularly if you are willing to sacrifice performance.

The simplest method is to make the whole job one transaction, so if it fails then NONE of the records are inserted and a rerun will process all records.

... cff.linkname_Sort,0: Fatal Error: Fork failed: Resource temporarily ... is the important message. This means that UNIX could not "fork" a process. Since this is integral to the operating system, it is not surprising that the jobs failed. Either you have a low system limit on nproc or ...

I can't think of a logical reverse-engineered logic for the selection with only 3 source rows to work with. I can think of several useless and unlikely formulae, including "the sum of each row and its predecessor must be even in Link1 and odd in Link2"...

Can you explain your logic to decide which row goes into which target link? Your example is missing that information.

Before executing the "dsjob" command did you source the DataStage environment by executing ". $DSHOME/dsenv"? This sets up not only your PATH but also your library search paths.

Different UserId/access rights? does "dsjob -lprojects" work?

I am a bit unclear as to what the problem is, apart from you getting different values from expected. Is the key value being reset to 38 on each run, or is it counting up appropriately but is now a low value? If you manually check your key file from a UNIX shell, what is the value there? How did you ...

Unlike the other COMPutational fields, the simple "COMP" is not standardized and depends upon the compiler writer. It is likely that we have a BCD encoding, so S9(9) COMP would take up 5 bytes. As fillers are usually just placeholders, and in your case contain the same value for each row d...