DSXchange

There is an event late in the job sequence's log called "summary of sequence run". This gives a list of the activities that were actually executed (omits those that were bypassed by the control logic or checkpoint logic).

For future reference note that address 0x80000000 is 2GB.

There is a sockets library in UniVerse, but it was added after the split from DataStage. There is no BASIC-callable sockets library for DataStage.

Merge stage will only show differences on columns defined as keys for the join.

Please advise which environment variable it was, so that others can more rapidly solve it if they get the same problem in future.

I am not aware of any environment variable that would affect the operation of mkdbfile, unless you were using that environment variable as a job parameter.

Have you deployed the STANCASS job?

No. Though multiple projects may, in some small way, reduce contention for locks held by client processes.

No, the problem here is that the command for CREATING the hashed file is not being generated correctly. Try re-compiling the job, then validating it.

A Sequencer with one input is a waste of space.

You probably need a Nested Condition activity ahead of Job activity 2; the test expression in the Nested Condition can determine whether the day name is Friday.

1. Yes 2. Yes, unless updating the same hashed file in the same job. In which case disable both write and read cache (since they're separate caches). If the hashed file is tiny (say less than 10 groups) I probably wouldn't bother with read cache either, as its pages are likely to be memory resident ...

The limit is in the data browser, not in the file or in the Sequential File stage. Don't worry about it. If you are concerned, use a text viewer of infinite width, such as UltraEdit.

Of course you can do in in server. Set up a stage variable for each comparison. Tedious, perhaps, but not difficult.

Parallel jobs have inbuild stage types for change detection (Difference, Compare and Change Capture stages). They tend to make this kind of task less tedious.

Try doing "deploy only" first.

Large projects are contraindicated because it takes longer to populate things like drop-down lists and tree structures in the clients. Small projects are contraindicated because of the need to manage shared components more closely.
Where are your priorities?

Why am I paying more taxes than the others?