DSXchange

You can set it to False if you like. What this ultimately means is that the optimizer can decide - if you are doing a direct copy with no changes particularly - not to include the Copy stage at all. After all, if it's a placeholder (a "do nothing" stage), then it's most efficiently replaced with a N...

If you export from Director there is (a checkbox). This does not seem to have been brought into either of the tools you use. Perhaps you could lodge an enhancement request?

I wonder what the "correct" answer (the one sought by the interviewer) is?

There is a small project in DS which is running fine till last week.

What has changed on this server? "Nothing" is not the correct answer, though the change need not have been in DataStage. Has "someone" changed NLSMODE from 1 to 0, for example?

Does it do what you require? I can assert that it's syntactically correct, but can not comment on its semantics.

And, indeed, what I had stated earlier in this thread.

Actually, you can even develop with the other versions, but you can neither compile nor run the developed parallel jobs unless you have version 7.5x2.

Parallel jobs are a new mindset. Your only option is to use job parameters.

These are set up as EQUATE declarations in JOBCONTROL.H, which you will find in DSINCLUDE on the server. All the status values have mnemonics beginning with "DSJS.".

... but here we're talking about server jobs, so your point is moot.

First off, get rid of the metadata mismatch. Import the Oracle table definition and load that (just the columns that you need) into your DataStage job. Even that will improve the speed a little. For the rest, the problem is what you are doing with the rows once they arrive. Chances are that the prob...

The answer to the original question is "extract the year and month from the timestamp". If your source is standards-compliant, and the timestamp conforms to ISO 8601, then the first seven characters of the timestamp contain the year and month. Because server jobs have no internal data types, it woul...

Just had to be sure. If you use Version Control for promotion all you have to is put the script (or a link to it) in any directory you have identified to Version Control as a "versioned folder" and Version Control will look after the migration of the script for you. You can "write the script in DS" ...

You need the UNIX-based ODBC driver for your database; this is probably the one that shipped with DataStage. Configure DSNs by editing a file called .odbc.ini (this is probably in the Engine directory for DataStage). If you're going to use that ODBC driver with DataStage too, you also need to config...

Is this a DataStage TX question (as implied by your use of the word "map")? If so, please post on the DataStage TX forum.

I believe I used the phrase "text file" when discussing the staging areas.

This is not the same as a hashed file.

The value of staging areas is that you can do the transformation phases and data preparation phases without needing to have any access at all to the source or target databases.