DSXchange

Sounds like your SQL is invalid. Can you post the *entire* statement?

dprasanth wrote:But I can do one thing, I can put all the records from the input file to a oracle table and then can join that with the lookup sql. I will try that and will let you know. Once again, thanks a lot!

Actually, that's exactly what I meant.

Edited the original post to (hopefully) be more clear.

As you've found, that can't be done with an OCI lookup as that option is only applicable when using ODBC or a UV stage. For a 'pure' Oracle solution, I prefer to load the 'source' information into a work table and join that to the lookup table(s) so that the 'multi-row result set' is what streams in...

Welcome. :D There really is no 'best' method, do whatever works and is maintainable by the crew you have there. I would think something of this nature would be better handled in a script, depending on exactly how much of the 'structure' of the file you are attempting to verify. However, like most th...

I don't believe this is a bug and just wanted to be sure you'd tried it with the Invocation Id. Honestly, this isn't something I've tried to do up to this point but since the Director can separate the log entries when viewing them, I'm sure it's possible, just not sure how. If I get a chance, I'll p...

Please provide more information. Details about your job design, your data, what you are sending to the field generating the error, etc etc.

I have refered to the available user manuals but could not find anything specific to AIX 5.3. You couldn't find anything specific to the 5.3 version of AIX or for AIX in general? The latter is found where you would expect it to be - in the Install and Upgrade Guide under the Installing on UNIX Syst...

I think 7.5.2 (usually referred to as 7.5X2) is meant only to run on windows platforms. No, 7.5.2 and 7.5x2 are two completely different versions. You are correct in that the 'x' version is Windows only and was 'special' in that it was the only version before the 8.x release that could run PX jobs ...

Basically... yes. The source can be anything, I typically land what I need in a flat file and then source from that flat file.

As Ray noted - using a single target database stage with multiple input links. Each link would populate one of the related tables and the link order would match any RI dependancies you have between the tables. In that situation, each 'set' of records can be committed individually if you enable Trans...

Ah... was going to point that out when I got a chance.

turash wrote:If transfer to each of the table is handled by an independent server job which is a part of a sequence job, can we manage the transactions across multiple server jobs?

No, no more than you could across multiple stand-alone procedural processes.

After you create a new 'Before/After' routine it will show up there.

You'd have to take this back to IBM, I would think, see if this is a known issue and if there are any patches to resolve it.

should we create a third layer on top of those for execution ease (another sequence that calls those 15), if it doesn't effect performance? Yes, you certainly can as long as it is valid to always run them together. And no, it won't affect performance any more than running the individual Sequence jo...