DSXchange

I would recommend that you set $APT_DUMP_SCORE on your job and you will most likely see that DataStage has inserted a repartitioning operator and/or a sort somewhere in your job and that this "hidden" stage is causing a slowdown. In the lookup stage how many rows and approximately how many...

I've never seen a job started with an invocation id fail to show it in the logs, so my thoughts on the problem are that there is some other cause that has yet to be found.

Does it compile and run if you manually type in the function? If so, then it was probably inadvertantly left out of the list in the file DSParams at that version.

If all of the decimals have the same behaviour plus your source is a sequential stage, you could do the conversion from string to decimal implicitly when reading the data and then specify handling for decimals in the stage attributes.

I'm not sure about the terminology here, since Job sequences don't have columns. Do you mean a sequence calling job or perhaps a job that you want to have call a sequence?

Unfortunately, I would say that you would need to code up a "IsValid()" for each of those conversions.

I assume that the OP's actual customer (the one with the support contract) is in NZ while the OP is offshore. It reminds me of the time I was waiting for my US-Visa for Ascential and I was doing a project in Canada; I need to call a special 900 number in order to arrange my interview and I couldn't ...

Yes, you need to define the custom routine in DataStage and also compile and bind your program outside of DataStage. I just took a quick look at the thread with the pxreplace code and it contains some instructions on the compile and links to threads, I suggest you start there.

Try using the "Search" button in the menu above, type in "Call DSExecute" as an exact search and -presto- you will see 46 matches, amongst which are several examples. Perhaps you could post your values for the 4 parameters. p.s. This isn't the correct forum for this question, but...

After a job is compiled it gets shown as status "Compiled" in the director log. Once that job is run, that status is changed (i.e. "running", "aborted", "finished", etc.). If the job is a multi-instance one then the main job name remains "Compiled" a...

It looks like you are compiling this as a server routine and it is interpreting this as a BASIC Program rather than a c++ program.

The re-compile was just suggested in order to track the problem and not to solve anything - you still have a job, somewhere, which is firing off the sequence without an instance name.

I have 8.7 here and don't see the stage.

The last post on this subject was 7 years ago and refers to a really old version of DataStage - hopefully this bug was fixed but perhaps the suggested solution might apply to this problem should it occur in a new version.