DSXchange

I forget exactly what we ended up doing as it was some time ago, but it involved editing the template that the script is generated from, from what I recall. You may be able to tweak it to echo out certain information or something you feel would help troubleshoot your issue, I would think. I was unde...

The biggest difference I would think, assuming wire drivers were used to connect via ODBC, would be the need to have an Oracle client installed on your DS server box. This may be an issue if your instances in question are always 'remote'. Other than that, a native driver would always be preferred to...

Not sure what you mean, Nick. The resolution was posted in the thread, and the generated script is supposed to be deleted after it runs in this version. I could probably post the contents and permissions of the generated script tomorrow from the office, still being on the older version and all, if y...

Not much - one example of a previous discussion on this topic here.

Since this doesn't seem to be the answer, what is the alternative? A query of DS_LICENSE returns as many rows as there are connections not licenses in use. Is there some quick way to get how many licenses it thinks are currently in use?

The point is, however, there's no need to do that. Almost everything the CFF stage needs to do that automagically is in the copybook, it just needs to be properly imported. Then I'm assuming that first field - HEADER-ID-MDBIM - can be leveraged to differentiate between header, detail and footer reco...

By first scanning your source file for the large table keys, bulk loading into work table in Oracle, and then inner-join the big table to the work table and spool to hash files, you reduce their volume to what works well with hash files. In all cases, only spool the columns necessary for your trans...

Check this post for some previous discussions on this issue.

David, I think your "Author Search" issue is simply because of your choice of a one letter user name.

2. Yes I could replace a few of the lookups with hash files - but we also have to access very large tables 30 Million rows+ so hash files can only be used in a few cases. The key is to only load into the hashed file the records you'll need for the current run. I doubt you need all 30 million each t...

ecclesr wrote:Where else should I start to look for where the bottleneck may be

Nowhere else.

Those Oracle lookups are the bottleneck. Why aren't you using hashed files? They are the preferred lookup mechanism for DataStage jobs, after all.

Should have mentioned that I saw that in another post and tried it earlier - it didn't change what was output. At least not for me.

Curious on how exactly I'm supposed to use the dslictool utility. Just had a situation where all 16 of our licenses were in use, so needed to find out who was on and who I could kick off. Querying DS_LICENSE gave me most of what I was looking for, but it wasn't obvious how many of the 16 DataStage t...

If there is a choice in the matter, then I whole-heartedly agree with you Vince. However, in my recent introduction to the Wonderful World of XML there was no choice given - feeds in and out to business partners like Yahoo and other search engines, build this huge zillion row XML file, read this hug...

When Ray chimes in on this thread, I wonder if the old roadblock is still in place. From what I recall, he has basically written a book on the subject but has constantly run afoul of Ascential's lawyers who have threatened to sue the pants off him if he publishes anything. Something like that, anywh...

And to me as well - was wondering if it was caused by an overly optimistic array size or some such.