DSXchange

I'm interested because I have a similar issue on one of the teams I work with and I'm wondering how best to address it. Remember that this function is sorting a delimited list of city names for each record processed - it is not sorting the file. Not sure why you would want to consider an external c...

Ah... what you should have done is specifically ask about 'Snapshot too old' issue instead of focusing on a perceived solution. Transaction Isolation won't help here. You'll need to talk to your DBA about the issue - they may need to increase the retention times on the UNDO logs, or advise you on h...

Could you please provide more descriptive problem?

What is 'not accepting' it? Handle_Null() must be used in a Modify stage from what I understand.

But that wasn't what you asked, but rather specified 'from the unix command line'. To call services from inside a DataStage job either as a source/target or row-by-row, you'd either need the Java PACK (if you wanted to roll your own solution) or the Web Services PACK, both of which would need to be ...

In DataStage the LEN function has not worked for us in this scenario. What exactly doesn't work about it? Can you provide some specific examples? Is NLS enabled or are you using something other than a single-byte character set? If more detail required please provide e-mail and we can send DataStage...

Yup, I get the same with Outlook. Ask your Admins if there is anything that can be done. I generally put something in the email body telling people to simply ignore it.

Not a DataStage question. Find out how to do that and then simply do the same thing from the tool.

Thanks Ernie. Interesting stuff, was blissfully unaware of the debates.

I must have missed Ray's edification request. I'll try to remember to post the combined wsdl when I get back in the office as others may find it helpful.

Suggest you visit Kim Duke's website and download the XMLPack Best Practices document he hosted there for me. Very thorough explanations on properly leveraging XML 'chunks'.

http://www.duke-consulting.com/DataStage_Tips.htm

Basically, this: MQ1->Transformer1->SequentialFile1->MQ2->Transformer2->SequentialFile2->MQ3->Transformer3->SequentialFile3 Perhaps read from a known empty file so that nothing is written to MQ but rather just serves as a process break... Still think they should be three separate jobs, control the o...

Interesting discussion here that might help:

http://www.dsxchange.com/viewtopic.php?t=104817

In other words, there's nothing obviously wrong here. So, perhaps the key values you are sending in don't exist in the table. Or, you've deleted them from one table and are checking in another.

Right, fine - as you write out the detail, send another link to the Aggregator to get your count, then write that as the header file.

Write to two seperate files, one header record and X detail records, then cat them together 'after job'.