DSXchange

A virtual Data Set is the structure. Data are buffered through them; it is not necessary to fill it before draining any rows from it. A Copy stage can be optimized out by the compiler. I suspect the support analyst was trying to get a feel for whether your partitioning algorithm was distributing row...

The warning is simply alerting you to the fact that some int32s have more than six digits. If you're prepared to ignore this information create a message handler to demote this particular warning to an informational message.

You may need to provide a format string for the timestamp. The default default for Timestamp is "%yyyy-%mm-%dd %hh:%nn:%ss" but your data seem to be "%mm/%dd/%yyyy %hh:%nn:%ss".

What columns have you identified as key? Have you identified any further comparison columns?

DataStage 390 generates COBOL and JCL which are run on a mainframe running OS/390. Do you have one of these? If not, your choice is already made. There is an extra charge for a DataStage 390 licence. Parallel and server job run on UNIX (or Windows) servers, and can access mainframe data through gate...

Writing your own plug-in stage is (theoretically at least) also an alternative. This is the reasoning behind the entire plug-in architecture of DataStage; to be able to create components to do things that the standard one's can not do. Is there an API for communicating with postgresql ? The manual P...

Yes, in theory. But they are free. They can also be given out during events such as the Server to Parallel Transition Lab. But ask your support provider first; let them earn some of your maintenance moneys!

Don't forget to ensure that the 32-bit libraries occur ahead of the 64-bit libraries in the shared library search path used by DataStage processes (unless you're on Tru64 or Itanium machines).

When you're number 10 you have to try harder. (Paraphrase of old car rental ad)

Do the append in an after-job subroutine. Call ExecDOS and execute the command type file2 >> file1. By the time an after-job subroutine is executed, all output links (and the files to which they refer) have been closed.

It would appear that something has gone badly awry in this project. I assume you are importing from a DataStage export file. My best advice would be to go into the Manager client and delete all "your" objects (not the ones supplied with DataStage) - jobs and routines particularly. Then re-import the...

If Char(N) was specified it was obviously for a reason (ok, or ignorance, or the default in the tool being used). It's in the spec, you have to follow what the spec says. Or challenge the spec, come to the right agreement, and get that decision recorded and signed off . This last step is commonly kn...

Cats(A.Col=B.Col ) = sv1 If IsNull(sv1) Then sv1 Else @NULL = sv2 It is not permissible to assign to the @NULL system variable as you have done. I suspect you are trying to indicate "assign to". Maybe ==> would have been a better symbol. What are you trying to achieve with the Cats() function? As y...

It is not possible to call a stored procedure from a Transformer stage (other than through a custom Routine, and even there it is not a good idea). So you don't understand completely what your job design is actually doing. A passive stage (for example ODBC or Stored Procedure stage) exposes the meth...