DSXchange

Indeed, DataStage will not be able to check that the number of columns in your design matches the number of arguments in the stored procedure, so is likely to generate a warning to this effect.

That's my third way, with all the coding work already done by Kim!

Almost nothing with two operands can be accomplished in the Modify stage. About the only exception is handle_null().

Substring is possible; concatenation is not. String length, raw length are possible.

Arithmetic using arithmetic operators is definitely not possible in a Modify stage.

kduke wrote:Good point. In the osh is there anything that identifies the job like link name?

No.

The best way, since you ask, would be to have a process metabroker. This captures process metadata automatically. The next best way is to rely on DataStage capturing the information into the job log, and retrieving it from there (no extra work re-opening the job). Each "active stage finishing" messa...

There's no such thing as a Lookup stage in server jobs. Are you in the correct forum? Even if you mean a Transformer stage, is doesn't mean that that's where the problem actually is. Active stages invoke methods that are exposed by passive stages; it is much more likely that a GetRowByKey() method h...

Two words: Process Metabroker

Automatic capture of process metadata.

No, because I am not in a position to monitor your hardware and software.

I just re-read your original post, in which you discovered that writes to hashed files (not "hash" files) are destructive overwrites when performed via a Hashed File stage. If you use a UniVerse stage, however, you are using SQL, and will need to insert (if the key does not already exist) or update ...

You can replace VOC pointers by adding OVERWRITING to the SETFILE command.

Otherwise you can delete from VOC using SQL:

This is an auto-committed transaction.

miwinter wrote:I've found the problem - thanks for that :D

We share here. Even if it makes us look temporarily foolish, it may help someone save some anguishing time in the future. What was the problem, and what was your solution?

I doubt that this will work; the [ ] indicate a substring operation, but the data type is date, not string.

Possibly, remember that the DB2 server is also competing for these scarce resources. Or the DB2 client software. Or both.

Welcome aboard. :D

The Merge stage exists precisely for this purpose.