DSXchange

Sequencer logic is easy. The output fires if "Any" of its inputs fires, or if "All" of its inputs fire. What you have there is a job sequence, not a sequencer. Your problem is that the job sequence is doing exactly what you have asked it to do, by checking "automatically handle failure" and not incl...

Probably not, since it's receiving its default value on the output. You want to have your cake and eat it too - to pass the column through without using it as a change column. That's not how Change Capture stage works.

Search the forum for "vertical pivot". It's do-able in several ways (though easier in a server job).

Search the forum for "vertical pivot". It's do-able in several ways (though easier in a server job).

NEVER use kill -9 on a DataStage process. There are too many interdependencies. If you must kill the process (which I doubt you should ever need to do) use a gentler method such as kill -15. This gives the process a grace time in which it can clear up resources.

Prove it. Post the pertinent sections of the score.

No.

When you upgrade to version 8.0 you get a new construct called a "parameter set", which will make life easier.

But, until then, the answer is no.

I never use RCP, for reasons associated with total metadata management, so someone else will have to contribute here.

Not supported by DataStage.

The Quote property issue is a known bug. If you open the stage properties you will find that "null" has actually been truncated to "nu". Use the GUI to change it to "none".

Fix one thing at a time.

Unless you have NLS enabled all the characters you are processing are ASCII. ASCII is a way of encoding characters encountered in business data processing. It is possible, of course, that you are receiving EBCDIC data (a different way of encoding), but in that case none of your characters will be AS...

HashedFile UniVerse (lookup) (user-SQL) : ^ : | V | HashedFile ---> TransformerStage ---> (any) (source) User-defined SQL in the UniVerse stage can effect the deletions you require.

Basically you are performing a lookup to determine whether the record already exists in the target table so the decision about new rows is already made (lookup fails). Then you effect some form of change detection - there are several approaches available - and determine whether there is a change. Th...

You/we can't tell where this is occurring, since your operators are being combined. But, at a guess, I'd say that somewhere in your job the EmpName column is defined as not nullable or that you are using a stage type that does not tolerate null. With no knowledge of your job design it is impossible ...

You can't do it in the Transformer, because you can't even SELECT the NTEXT data type. The extraction SQL could, of course, CAST as NVARCHAR or NLONGVARCHAR. I've also seen an approach where the extraction SQL brought in "chunks" (substrings) of a very large text string - from memory four columns ea...