DSXchange

Yes, it means your execute command activity finished
Need more info on that. What command is it trying to execute. What surrounding error messages are you getting?

gpbarsky, you need to bring your input into plain numeric format before using ArndW's code. So all together your expression will look like this

If you are well versed in shell scripting you can write your own script. Restartablilty can be handled by reading the job names from a file and running them sequentially, once any job fails, copy the rest of the job names from the file to another file. Go investigate the error fix it. Next time you ...

I think I know what kjaouhari meant.
That instead of going through the hashed file, is there any other way to find out inserts.
Right?

Need to look at your routine. From the looks of it, you did not let the compiler know that you will be using other user-defined routines within yours. You need a deffun prototype at the top of your routine.

What is the warning message that you are getting?

DSguru2B wrote:Secondly, is tmpLoc an environment variable?, if not and its just defined as a job parameter then get rid of $

You did not answer the second question i asked?

First of all, is your datastage server on unix or windows. your heading says its windows and your trying to run a unix command. Even the directory structure is of unix (the forward slash in #$tmpLoc#_data/LoadData.dat) Secondly, is tmpLoc an environment variable?, if not and its just defined as a jo...

How are you scheduling your jobs?
Are you using an entreprise scheduler like Control-M or Tivoli, or by cron tab ?
Regardless, it should not behave like this. Are you sure that your providing the correct scheduling requirements with proper scripting, if its involved.

When inserting the universe table, did you identify the records with some kind of source code to identify its origin?
If not then i dont think you can. I might be wrong, but as far as i know, no way.

Check out Ray's reply in thispost.

Try loading to a sequential file first and see whether , really, you got the conversion successful. Most of the time in such situations, the blank space between the date part and time part of the timestamp is missing. Also, if you post the derivation, it would help us pin point the mistake, if there...

You must be derive the rejected columns from the source and not the hashed file. I bet you are deriving it from the hashed file. The lookup failed and hence the values will be NULL. Makes sense?

Do an exact search on "Error occurred during link open processing". It reveals a few posts. See if they help.

Excuse my ignorance, but doesnt Oracle take in the timestamp as
YYYY-MM-DD ?