DSXchange

From a Server job.

Done.

You load B, C and D into individual hashed files, stream in A and then join to each via a reference lookup one after the other. You can technically do this all in one transformer but beginners should stick with three transformers with one lookup apiece. Add the columns you need from each lookup into...

Unfortunately, that still doesn't give us any diagnostic information to work with. Did you try the 'Reset' I mentioned? Have you tried running this same SQL insert in your tool of choice (i.e. outside of DataStage) to see if it works there or if you get a better error message there?

Are you looking to get this "30 days job status information" from something stored in Autosys directly or do you just mean you use Autosys to run your DataStage jobs and you are looking for the last 30 days of information from the DataStage logs? :? Something like ETLStats will help with t...

Can you post the details of your job design, please? And you get this error when, exactly? Runtime?

Typically one would "source" the dsenv file in order to set up their environment properly, which would include the items rameshrr3 mentioned.

I did this years ago and recall that there was a system variable or something to check for end of file but I believe that was in Server, not PX. There's no such beast in either product, something they will allegedly address in the 'next release'. Now that 8.5 is out, I wonder if that actually made ...

Any suggestion at this point would be a guess as there's not enough information to do otherwise. Next time your job aborts, Reset it from the Director and post the contents of any "From previous run..." log message that appears.

Not sure Kim has any control over that. AFAIK it's up to your browser, its settings, how you click on the link (shift/right) etc.

OK, let's start with a simple solution and then work from there. Why not continue to do the join in SQL, all at once in a single stage? While you can bring in each table individually and then let the job do all the work, there's no reason to go to all that trouble and complexity unless you're being ...

Nobody said you would have to pull the data twice. Can you not land the result set and then check that result set is not empty? Doesn't have to be a flat file and there is (off the top of my head) an orchadmin command to get the count from a data set. I honestly don't know of any way in a single job...

So, yes to the last question. Then you'll need a Sequence to check the source and then conditionally run the load job only when there is actually data to load.

Nice! Don't forget the secret handshake.

You can use whatever you like, but that choice will drive how you check it... i.e. how you get the data back out. A simple 'cat/type' command won't work with anything other than a text file.