DSXchange

Need more information. What is the record schema? Where are you applying the left_substring() function?

Do you have a Custom stage or a Wrapped stage? It's the program itself (the "UNIX command") that failed, so you'll need to concentrate your diagnostic efforts on that.

Or you can be selective during import. "Import Selected".

You don't need a UNIX command - the Status statement in DataStage BASIC can return the size of a file in bytes.

While it is always problematic to react to a single sample, I'd essay that spending 64.5% of the time in kernel processes suggests that something is very wrong indeed - perhaps the machine is thrashing. Get an operating system specialist to take a look at it. Maybe there are just too many competing ...

Most IBM Information Server training is delivered on a VMware image.

Yes, because DSXchange is about the Information Server suite, and we need to know (possibly) which job type people are having problems with. The version can be important too, especially as more seamlessness comes in with versions 8.2 and 9.0.

Rather than that, Search DSXchange for the SETFILE command, which is used to create VOC pointers to hashed files.

The conductor node is the machine on which the conductor process runs. Usually it is the first node mentioned in the configuration file provided that that node is in the default node pool. The conductor process is the first one started, composes the score, requests one section leader to start for ea...

Did you install the compiler before you installed DataStage, as required?

Code -3 can only mean what is says - that the parameter name used in the DSSetParam() call is a name that does not exist in the attached job. It may be a mis-spelling, it may be incorrectly cased, it may not exist at all (are you sure you have the right job attached?). You need to check thoroughly.

If you imported from the COBOL File Definition and used that "table definition" in the CFF stage, then the only "extra" thing you may need to do is to indicate in the CFF stage properties whether to normalize or to flatten the array.

samplify wrote:(oh yes and i need that many multiple rows being written to the output)

I still don't understand why (that is, how any such output could be useful/meaningful). Maybe it's just me.

Or read the chapter on the CFF stage in the Parallel Job Developer's Guide - I recall, whether accurately or not, that only one OCCURS clause is supported in version 7.x.

Have you set up any DataStage credentials in Information Server?