DSXchange

You need other stage variables for the other columns. Please try to generalize my example.

Parallel transformer functions are also intolerant of null values in other arguments. That's why I asked (twice) about the input columns.

Incidentally, "" and NULL are not the same thing.

The Input Value field contains MyJobName MyStageName MyLinkName Within the after-job (or after-stage) subroutine you use Field() functions to parse these out from the single string. JobName = Field(InputArg, " ", 1) StageName = Field(InputArg, " ", 2) LinkName = F...

Can either of the input columns ever be NULL?

I don't think you can do that either, but you CAN pass an entire SQL statement as a job parameter, so that may be a viable workaround for you.

What happens when you open a cmd shell and type the AT command?

Use a Column Generator stage to generate sequential integer keys on each link (outputs from the source files), and use that as the join key.

Can either of the input columns ever be NULL?

It is much faster to delete a project using the Administrator client. There are too many different things to delete if you try to delete from the command line.

Check that the disk is not full where your project is installed. Check, also, that your log file has not reached 2GB. Either of these situations will cause the corruption that you see. If you don't need the log contents, you can clear the log table using a CLEAR.FILE command (search the Forum for de...

The BEGIN and END tags are characteristic of DSX export files. Can you please try downloading it again? Others have had success with it. You need to Import DataStage Components to get the routine into your project.

The answer, definitely, is NO.

To what values were your NLS_LANG and LANG_C environment variables set?

It's in the Sort stage.

Another way, particularly if the headers and trailers have static text, is before/after subroutines. In the before-job subroutine you echo the two header lines into the file. echo Header1 > #filename# echo Header2 >> #filename# In the job itself you append to #filename# In the after-job subroutine y...