DSXchange

OK, I take that back. FIVE times now, why is this so hard for you to get right? One post in the proper forum is all it takes. :roll: This is not a metadata question. It is a technique question. It should be posted in the forum appropriate to job type (server or parallel). Metadata Workbench is not D...

Dude, you've now posted this same question FOUR TIMES in the wrong damn forum. Post it in the proper forum based on the job type and people will be more than happy to share helpful thoughts with you.

You just need to create them empty? Empty empty? It's not like a database table where you create it based on metadata, for that on UNIX just use 'touch' and the filename to create the zero byte placeholder. If you need a header record with column names, you could cat that to each name to create a 1 ...

Investigate the use of the DaysFromDateSince function, both by searching here and by reading the documentation.

Where is your expertise - Server or Parallel? What part of this do you need help with?

Generally, one overcomes problems by answering the questions the people trying to help have asked.

DB2 does that. Make sure your update or delete actually has a row to update/delete before you fire it. Or use a message handler to demote the warning to informational.

Which part is unclear? Basically, all I was noting is what we are doing - building the XML files with what is in essence a static name and then renaming them afterwords to what we really want them to be. Part of the process writes out to a flat file (from the data) the information we need to do the ...

I don't think kamalshil noticed not just the order was 'scrambled' but the first and last names were as well. What you are doing is called data masking and it is not all that easy without dedicated routines / functions. IBM does supply them, but (for now) only through engaging their services. Some d...

This is not a metadata question. It is a technique question. It should be posted in the forum appropriate to job type (server or parallel). Metadata Workbench is not DataStage. This forum is for Metadata Workbench questions.

Wherever you found it should have included some information about its use, I would think.

Welcome. That's not really a DataStage question, though some here may be able to answer it. You are probably better off posting in an official Teradata support forum or pinging their support people for help, especially if no-one else chimes in here.

You mean like 'F' and 'U'?

Then simpy set the "Abort after X rows" value on the reject link to 1 and the first record down it will abort the job. Now, someone else will need to speak to how this is impacted by running in a multi-node configuration in a Parallel job but I believe this should work for you.

The answer will be different, depending on if you want to track all rejects and then abort after all have been logged or simply abort on the first one you find.