DSXchange

The third log entry ("Environment Variable Settings: (...)")in your director for a job run will contain a list of the environment settings for a given run. Can you see DSHOME and APT_RESPATH?

I wonder if when you executed the "DS.REINDEX ALL" command that you may have had DataStage users logged in and thus invalidated certain indices, which might be causing your problem. Can you ensure that nobody else is using DS and re-execute that command?

... Please it is very urgent. is a guaranteed method of getting a slower response at DSXChange; it is a volunteer site and most of us are being paid for doing "urgent" work. I think your problem is sufficiently different as to make a new thread necessary, as it would seem to involve TeraD...

The file is also a text file called "uvconfig" in UNIX. You posted the comments from the file, but not the actual assignation statement in your post.

I am really surprised that you had no success, since this topic has often been raised and answered in threads over the years. I went into search and used "64 bit hashed files" and of the 16 threads found I suggest you look into this one to find what you want. If you need to create the file...

This might be a good time to restart the DataStage service completely or even to re-boot the DataStage server machine. This would, among other things, get rid of any locks that might still be around and could be part of your problem.

It sounds like you are only using one sequence name for both, you would need to have 2 separate ones, one for cust_id and one for merchant_id.

Check the text file $DSHOME/uvconfig for the file setting and, as a cross-check for the actual value used in the running version of DS enter "$DSHOME/bin/smat -a | grep UVTEMP". These two should be the same.

Specifically for Oracle you should ensure that NLS_LANG is set to the correct character set.

You don't really need to do much, if anything, in DataStage. Tell DS that the source file is UTF-8 and ensure that your environment Oracle variables are correctly entered and the data will load correctly (the "view data" won't show correctly, but that is a known restriction in DataStage)

The first thing you need to do is to get some solution for just one job that works. Once that is done you can figure out a way to generalize that solution for the whole project - this might involve doing a project export into a .dsx file and a global replace or perhaps even more involved manipulatio...

There are two basic approaches to this - either add a trigger and call a stored procedure which uses SQL to update the previous and next records in the table, or to write the same logic in DataStage. If you choose DataStage to do this you once again have a couple of possible ways to do this. One met...

Can you check the crontab contents with "crontab -l" when logged into UNIX as the same userid and post that information?

Did you also mark the column as using "extended", if not then the microseconds would get truncated even if the format correctly parsed the timestamp.

Can you echo the parameter value in the job (i.e. to a PEEK stage) to see if the backslash characters are still in the string and that the single quotes have not been stripped?