DSXchange

But, I do not want to use the Hash file as tab1 is very big and there are very few input (say 50-60) compared to it. Is there really a need to load all of tab1 into the hash? As Ray noted, 3 lakhs is not all that big - however - you are correct in the fact that you don't need all of them in the has...

Here are a couple of different ways. Not sure you can from dsjob, however.

You may also find that this is a Server only pack and not really applicable to Parallel jobs, with the only problem being the fact that it is letting you try to put the stages on the Parallel canvas. I tried poking around on Ascential's site, but couldn't find anything definitive one way or another....

Not really a DataStage topic. A quick Google should turn up all kinds of things on Samba, including the home page. Here's a small quote from the What the Heck is Samba? page: Samba is software that can be run on a platform other than Microsoft Windows, for example, UNIX, Linux, IBM System 390, OpenV...

is it necessary to do trimming for both source and hashed file keys. You tell us, we can't see your data. You need to do whatever is necessary to ensure that the data matches across all keys - or, as you've found, the lookup will fail. If that means trimming any character strings used in key fields...

What might be the problem. Well, how about all the normal stuff... * Verify that the the keys you've chosen for the hash are appropriate * Verify that you are mapping the right input source fields to their matching hash keys * Verify there really is matching data between source and hash * Verify th...

Yes. It's called, oddly enough, substring and is documented in the Modify Stage section of the Parallel Job Developer's Guide.

Odd. No direct ideas, but curious about something... You keep saying 'I'. Does this mean it only happens to you and not to other team members on their client machines? If there's any indication that it might be certain machines, the first thing I'd try is reinstalling the Client software. Couldn't h...

And perhaps even read and write data to Oracle now.

Cool.

Me too.

I'd probably write all of your functional keys from the source DB2 table to a flat file during the run, along with a 'P' or an 'E' based on the success or failure of your Oracle work. A second job could then take the file of keys, source those and perform updates on those records back in DB2 with th...

Yah, I was considering answering the next batch of questions like this completely wrong, just to be a poopie head. If it was one of those Poopie Head Days, that is.

Keep in mind that the default sort is a string sort:

2 b 1000
3 c 20
4 d 230
1 a 500

You have to either use the -n option to get a numeric sort or right justify any number field you are sorting on and pad them with zeroes.

Make it work from the UNIX command line first and then it should transfer into the Filter and work just fine there. Be aware however of a bug in the Filter stage where certain characters cause the command to be truncated from the point. In my case, use of quotes or a pipe will trigger the bug. One s...

I don't have this information off the top of my head, but there must be a field that related DS_JOBOBJECTS back to DS_JOBS. Let's make one up and call it JOBID. Include that in the query: SELECT OLETYPE,JOBID,COUNT(*) from DS_JOBOBJECTS GROUP BY OLETYPE,JOBID I'm sure someone here knows the ...