DSXchange

I'm not quite sure what type of deadlocks you are looking at reproducing. The DataStage internal locks for such items as jobs and other repository objects are what are cleared by the deadlock daemn. Such a deadlock can be easily created by starting a designer session, seeing which UNIX process is be...

Welcome to DSXchange, Madalaja. Do you think that your question is more scripting or DataStage related? If you feel it really is DataStage, then posting the short-form of the script (no contents, just the parameters that you use) might help; but my feeling is that your question relates more to param...

"grep" is most certainly part of the MKS Toolkit. I'm really surprised that the command wasn't found, since if the $PATH isn't set correctly for MKS toolkit then DataStage itself wouldn't run. If you search your server for "grep.exe" to find the correct directory you can then mak...

The MKS Toolkit gets installed as part of DataStage and that contains quite a few standard UNIX commands. Thus the answer to your question is "maybe". Do you know which commands you need to issue?

While checksum algorithms are pretty efficient, they still take CPU cycles. I would put the CRC value in the data with each record, the will at least halve the number of computations if each record is only compared one time. Also remember that each field you concatenate needs to be checked for NULLs...

I'm not at a place where I can check, but I've done this type of log information several times before and it has always been a bit work to get functioning correctly. The code you've posted looks good and you haven't stated which part isn't working - I would guess that everything but the innermost Ge...

Your analysis of the problem cause is correct. While datasets are very efficient for reading/writing data, they remain at heart sequential files (perhaps "sequential files on steroids" might be more fitting). They are not databases and, along with sequential files, are 1-Writer or N-Reader...

Craig - I realized that after I'd typed in my message and thought I'd discarded what I'd written instaed of posting it. Oopsie.

srinivasan.v, that is a difficult request to fulfil, since DataStage is just a tool and your design isn't going to be about DataStage but about your specific problem and how you want to go about solving it (using DataStage). You won't get much in the way of answers to your question here, but perhaps...

There is no facility to mail to the forum or to a thread. What is easy is to post screenshots or documents to a hosting website and then refer to the URL here. Photobucket.com is a free website for hosting images and many use that to host screenshots.

Since datamine has opened as a new thread *here* we'll close this one and continue on the new thread.

vamsi.4a6 - The unicode property is part of the text columns, each one can have its own setting.
ntr - A hex display utility is some program that shows the hex values for data. A simple one one is "od -x" in UNIX and many text editors have a hex display mode

I can't think of a viable means of doing this efficiently in PX without resorting to writing a bit of code, unfortunately. If doing it in server without resorting to hashed files a routine that stores the records between calls (using a COMMON block) would work quite well and not be too complicated t...

Pratyusha, for exactly this reason you always need to ensure that your data is partitioned on the key(s) that you wish to aggregate on (or join, or sort, or any other stage that process grouped data). If your data is partitioned on a hash of Column "A", then you are going to get incorrect ...

Thanks for that. If you only insert/update 2 records do you get this problem or does it happen after many records have been written?
I'm a bit curious about the corrupted key since DataStage goes through Oracle APIs and any corruption points to an Oracle issue and not one within DataStage.