DSXchange

Investigate how well this "complex join" is supported by indexes.

Also make sure that it does not need to materialize unnecessarily large views before joining those; always work with, and constrain as much as possible, base tables.

Windows operating systems that are based in NT are perfectly happy to accept forward slashes in pathnames.

This includes all the Windows operating systems that are supported for DataStage servers.

Try a Search around the forum. The first question is which jobs you really need to convert, and which will run perfectly well - maybe even better - as server jobs. There is NO automatic mechanism. You can not, as you seem to be asking, export as server and import as parallel. That would be a really ...

Welcome aboard. Load the max date into a Lookup File Set with Entire partitioning. This has a constant (say "X") as its key, and the max date as a non-key field. Use a Lookup stage to lookup the record whose key is "X" from the Lookup File Set. You need a Column Generator stage upstream of the Looku...

StringToTime() requires that the format string show what the string looks like, not what you want the time to look like. StringToTime( DSLink.Input_col, "%h%nn%ss") You probably need to be a bit more careful, with an If..Then..Else testing whether the length of the string is 5 or 6...

If you manage your table definitions in the Repository, like you ought to, then it's simply a matter of clicking the Load button or dragging the table definition onto the link. What could be easier?

There are DataStage macros that will return these values. You can also use DSGetJobInfo() function in a routine, or the dsjob -jobinfo command from the operating system.

I suspect that the problem is more likely related to the "#" character in your invocation ID. Can you try it with a different invocation ID, one that does not have a "#" character?

The Data Set descriptor file is not interesting - all it contains is the record schema and the locations of the actual data. Why do you want to read that?

On the other hand, repairing the damage is a potential source of revenue for consultants such as KBA - on that basis you could understand Ken's glee.

You could create a Routine activity from a job sequence that returned the date/time created/modified/accessed for a file. Use OpenSeq, Status and CloseSeq statements. Or use an Execute Command activity and the appropriate command for your particular operating system.

How about using the RMM stage?

There is also a "certification preparation" class offered by IBM, but this is usually offered only in conjunction with the Information on Demand conference.

The Join keys must not only have the same name, they must also have the same data type.