DSXchange

Basically create a separate join for each circumstance and develop logic to determine which result to use.

Just one folder is certainly possible. Create an export of just that folder's contents.

That's how DataStage works, and always has. It adds leading and trailing nonsignificant zeroes to the extent specified by the Precision and Scale settings. Don't worry, it's still the same number and will produce the same results in arithmetic. You can only remove the zeroes if the target data type ...

Check the Transformer stage properties to determine whether you have legacy null handling enabled.

You have to "create" the virtual table definition first - by specifying the columns and WHERE clause and giving it a name. Once you have saved this it will appear in the list of tables available for running Column Analysis (and, later, the other analyses).

External routines need to be written in C++, not C.

You can use a sequence. Try using job activities without links, to see whether they wait - the sequence is supposed to start as many jobs as are eligible and wait for any of them to finish. I just checked this on a version 8.5 system and, indeed, that's what happens - three jobs running simultaneous...

You should have read access to the install log. The product should have been installed by root with umask 002.

In the sequence job properties do you have automatic handling of job failures enabled? Try disabling it.

There are two sets of child jobs; design-time (the job activities in the sequence) and run-time (the jobs that were actually run). How to get these has been discussed in the past on DSXchange so you should be able to find them using a Search.

The problem here is in the operating system, not in DataStage. You are not permitted to have multiple open-for-write calls at the same time for the same file. Run the three outputs into three separate Sequential File stages populating separate files. Use a condition to detect month number and rename...

For a left outer join all columns from the right input may be null and therefore must be nullable.

To be transferred to the output without default values being invoked, those output columns must also be nullable.

If that wasn't enforced in version 8.1 it should have been.

Are all columns on the right input (as well as output columns derived from them) set nullable?

What is the actual exit status (e.g. from Director) of Job1 ?

In your scenario, if Job1 finishes with any status other than the two explicitly mentioned, Job2 will not be triggered. This includes not only DSJS.RUNOK (which is your intention) but also any other status such as DSJS.CRASHED.