DSXchange

Each job activity should have a Custom trigger firing on finished OK or finished with warnings.

Perform a lookup against the target table. If the row exists it's an update; if the row does not exist it's an insert. You may find it more efficient to perform the lookup against a hashed file containing the keys from the target table.

No.

You could always create one. More arguments, more variables in a COMMON area.

Execute the following command on the server (not on the client)
Post the results.

Post a picture of the Job tab in the Job activity so we can see exactly what you are attempting.

Welcome aboard. :D

Search can be your friend. The command line interface for running DataStage jobs is dsjob

Michael's contribution does hold true in most, if not all, databases. This is usually because common practice is to index foreign key columns.

Congratulations for solving. Now you can mark the post as resolved.

Search can be your friend. You are not the first to have had this requirement.

There are no books. The two main web sites for developers are this one and IBM DeveloperWorks IBM also have sales information, product announcements, etc. Other than that, fire up your favourite search engine - DataStage is not a name that will get too many false positives. You might also like to le...

Set NLSMODE to 0 in $DSHOME/uvconfig using any text editor.
Make sure nothing is happening. Stop DataStage. uv -admin -stop
Regenerate the shared memory segment. cd $DSHOME ; bin/uvregen
Start DataStage. uv -admin -start

Surely the Change Capture stage requires identical key partitioning (so that valid matches can be made) AND identical sorting (so that memory can be efficiently used). Try adding sorting on both input links, using the key column(s) as the one(s) to be sorted. That way, out of order rows will not be ...

Continue to search back in the code, to find out what IdV0S27%%UserVar%%4 through IdV0S27%%UserVar%%7 map to - they should be variables in your User Variables activity. It is clear from the code already posted that the generated code appears to assign these properly to the p$... variables that are u...

Lookup stage is to be preferred if the reference data set will fit in memory. It does not "reduce the performance". And the inputs do not need to be sorted. Partitioning must ensure that every valid lookup will succeed. Join stage does not support reject link. But you could employ a Join stage with ...

This indicates that the @SELECT phrase, or @ phrase if you do not have an @SELECT phrase, in the dictionary does not refer to Department_Number. It has to be there in order for it to be accessible to SQL queries. @SELECT has the task, officially, of describing what fields will replace "*" (so that d...