DSXchange

Search the forum for discussions on SETFILE and the UV/ODBC stage.

My guess is that you were trying to import metadata, and the query to get the list of tables failed.

Inability to do this usually means that your userid lacks the appropriate permissions in your database to select from certain system tables. I suggest you pursue your DBA and get the permissions.

CoolHard wrote:I am very appreciated if I get reply immediately.

What???

I would like all of my questions answered immediately as well.

Search the forum for macros, or consider writing out a standard file name and then using an after-routine to rename the file. That gives you a powerful option.

No

Here's a shell script you can read to learn about the dsjob executable:

http://www.dsxchange.com/viewtopic.php?t=85578

Nic wrote:I always use Administrator. On the Projects tab click on your project and use Command. On the command line use:

Code: Select all

delete.file hashfilename

This should be the safest option.

This only works if the hash file was created in the project using the default option, not an externally pathed file.

Client is always Windoze, release 7.5 something works on Windoze and Unix.

As long as you didn't use "SETFILE" or your hash file was created by a job using a project name or the default "account" option, you are safe to issue the rm command to remove hash files. There is a "D_filename" file in the directory where the hash file resides. It's easiest to issue an "rm -r *file...

I assume you mean hardware requirements, not OS. Hardware is subjective. How much data, how often, how quick? If the answer is gigs hourly in under an hour, then you are talking differently than megs each day between midnight an open of business. If your question is absolute minimum to just move a s...

Yeah sucks, don't it? You have 2 mechanisms for aborting a job once a link reaches a certain row count: 1. Hard-code the abort after rows limit in the constraint of a transformer stage output link 2. Use a polling type job control that periodically uses the API call for getting a links row count dur...

Use separate output links and write to separate sequential files using the appropriate metadata for each file. If you need the rows collated back into some semblance of order, add a grouping key column and send the files thru a Unix sort command collating and ordering the files appropriately. DataSt...

Go back to your original post and read the answer supplied there.


ETIQUETTE: Do NOT multiple post the same question.

One

Possible solutions: 1. Write a job to spool output to a text file. Write a ksh to parse and create the appropriate directories. 2. Write a job to spool output to a text file. Write DS Batch job to read the text file, parse it, and issue Unix commands via calls to DSExecute to create the approriate d...