DSXchange

The expression tkey=1 evaluates to 0 unless the value of tkey is 1.

Try

If you haven't created the other job yet, you can use File > Save As. There are no utilities within the product for copying job properties between jobs. Stages you can copy and paste (in the same client), but job properties you can not. That's probably the strongest argument for using Templates at t...

THE difference?

A server job runs on a single server.

A parallel job runs multiple, parallel execution streams. This can be on a single (multi-CPU) server or on multiple servers.

Are you using the same user ID in your DB2 stage? If not, does THAT user have SELECT privilege to syscat.columns ?

Do you have a leading space (reserved for the sign)? Check with a Peek stage or similar.

That is a really messy requirement, whether in SQL or DataStage. If you can think of a way to do it in SQL, we would have less trouble finding a DataStage solution.

I'm confident there is no solution available using the Merge stage.

Do you have SELECT privilege on the system table in which column definitions are stored?

In each database there are constants for current date and current time. Their names vary. I don't have the DB2 ones in my head. Names like SYSDATE or CURRENT_TIME are the kind of thing I have in mind. But I don't have enough (unpaid) time to research them for you.

Even if you get it to work, DSStopJob leaves your job with a status of Stopped, which means it needs to be reset before it can be run again. This is not good practice. Much better would be to perform the initialization in (from) a job sequence and, on the basis of that initialization, make the decis...

ODBC data source names must be created in the $ODBCINI file (usually .odbc.ini ) but that's not enough. There must also be a reference to the DSN in your uvodbc.config file (there is usually one per project) indicating that this name is an ODBC - rather than UniVerse - data source. ODBC for SELECT p...

A pipe, by its very nature, expects to have its input and output open at the same time.

When not using a pipe, the Sequential File stage - being a passive stage - can not open its output until its input has closed.

Depends on the row size and what size of internal pointer. Therefore an answer in terms of rows (records) is not practicable. With 32-bit pointers, a hashed file is limited to 2GB of data. With 64-bit pointers, a hashed file is limited theoretically to approximately 19 million TB (though in practice...

Reset the job in Director. You will probably get an event logged "from previous run..." containing diagnostic information. My guess would be (using Oracle terminology) that you are sending the entire load as a single transaction, and the size of the rollback segment has been exceeded. If that's the ...

You said identical, but please check. Clearly something is different. Is write cache enabled for both? Is write cache size the same on both systems? Is the total of other activity the same on both systems? Are the hashed file configurations the same on both systems (in particular is MINIMUM.MODULUS ...

The UniVerse/DataStage shell (uvsh or dssh) doesn't really support the concept of environment variables, which is why an indirect mechanism is needed to get at the parent shell's environment variables.