DSXchange

You need a mail application installed on the DataStage server which needs to have a command line interface.

Adding just a call to DSLogInfo() will not compile in the routine, I meant adding something like CALL DSLogInfo('Arg1 is "':Arg1:'", Arg2 is "':Arg2:'" and return value is "':Ans:'".','') and seeing what the return value actually is, since it is neither 0 nor 1.

malloc() is a pretty fast and efficient call; plus the space is release after the call is complete so no fragmentation occurs due to these calls. I'm not sure if the implementation of fast pooled memory allocation are portable across UNIX versions, as the calls seem to be different. It would certain...

There is no "reset if required" from the command line, but you can always just force a reset and then run the job.

That shouldn't happen, the routine normally has no idea of where it is being called from.
Have you added some CALL DSLogInfo() calls to your routine to make sure your input and return values are what you expect for each call?

DSGuru2b - if you have multiple engines running you need to use the user/password/server command line options, or if the user you are calling dsjob from doesn't have DS permissions; but apart from that you are correct in that those options aren't required.

I haven't seen that problem in this forum, but recall having had issues in another forum using the same phpBB product. I cannot recall what changed in order to make the problem go away, though.

Which browser / OS combination are you using?

...not using all available memory and causing high page faults... A page fault occurs when a page requested by a program is not in memory and needs to be fetched from elsewhere. Often this is corrupted and misinterpreted to mean that the system is paging (particularly in Windows environments). Note...

Perhaps these interview question answers might help those who decide to do DataStage work instead of selling real estate

Is drive F a local or remote drive? It looks like you are creating a hashed file and that path on F doesn't look good with a ".|" at the end of the message. Can you check your hashed file stage for a valid file path and name?

In this case I would prefer using simple UNIX commands to put together your .ctl and .eot files instead of writing PX jobs or using DS BASIC code to do it. It is simpler and easier to maintain that way.

And is the PX system creating this customer.dat file or is it coming from somewhere else? If from somewhere else, how are you going to get the information about "time of file creation" and how do you define end-of-transaction?

Are you creating this "customer.dat" file in your DataStage job? Also, is this Server (as indicated by forum) or Parallel (as indicated by post).

In Server jobs you can capture failed writes to an OCI stage by adding a transform stage set before your output. In that stage you can use constraints to a reject or other output link and then choose what you wish to do with the rejected rows.

-username is an invalid switch, use -user