DSXchange

Do you mean the maximum number of concurrently managed sequences (unlimited) or the maximum key value that can be generated in any particular sequence (99,999,999,999,999) ? The latter could be higher if the configuration parameter EXACTNUMERIC were changed.

Have you Searched ? It's usually a symptom of an overloaded server.

Are they asking about design-time architecture, compile-time architecture or execution (run) time architecture? Each is different.

Is there some reason to have the "t" bit set on /tmp? What happens if you chmod -R 777 /tmp?

Yes, but it must be version 7.5x2, which is the only version that has job compilation and execution capability for the Windows platform. With other versions you may be able to create parallel jobs, but you will not be able to compile or to execute them. Note, too, that you will need a C++ compiler; ...

The warnings appear because there are situations about which you need to know. You can not suppress these warnings; you can only prevent them, by using your ETL logic to guarantee that no row is sent to the target table that will violate any constraint on that table.

If Director has folders, then the respository has Categories. They have, in fact, been present since version 1.0, though with a maximum of two levels initially.

From memory a three digit decimal number (such as 111) or a four digit hex number (such as 00f6) ought to work. You may have uncovered a bug. What does your support provider say?

In the interim I can only suggest that you edit the generated JCL before sending it to the mainframe.

This is an error within the client software itself. It usually means that a number (counter?) within the client application has become too large, or an attempt has been made to access an element larger than dimensioned in an array. Do you have a very large number of elements (jobs) in that particula...

Nor did you mention exactly what error you are getting. It is helpful for diagnosis if you post the entire error messages.

A good way to remember is that, as far as the database server is concerned, a DataStage job is just another client application. Therefore it (DataStage) needs to have the client software installed.

If you have EE you do have the Enterprise stages.

Loop Remove JobName From Output Setting MoreJobNames * do something with job name While MoreJobNames Repeat This is slightly more efficient than a counted loop, since the DCount() function must process the list once then the counted loop must process it again. This uncounted loop only processes the...

... provided you can guarantee that the first row will go along the first output link!

How about "the no. of columns in the source file will become the no. of rows in the target file" in the original post? That sure looks like a pivot requirement to me. Prior to the Pivot stage you can introduce a new column that contains the field names (which become column 1 in your output). Derive ...