DSXchange

Show us your input string, the derivation you are using, the output you are getting and what you want it to be.

Look up the "Target/Write Method" property in the Parallel Job Developer's Guide or the online help. It allows you to specify a field as a "Key" and then use that key value in the filename.

Hopefully your target field is a string. True DATE data types are just an internal number and thus have no format so I do believe you can skip the IConv and try it with the OConv alone. Either that or leverage whatever DB2 uses for a "to char" function directly in the source SQL.

I don't believe there is a DataStage component to this and thus no use of dsjob... best I can tell, anywho.

Hmmm... mod? Mod by 10 and then add the remainer? Decimal values you may need to round up first, me thinks.

Use Index() to find the correct dash and then use the position it returns in your substring.

Any update on this?

Why not try it and see? Assuming it's a contiguous number, your mod() mechanism would distribute them fairly evenly. However - the results of the mod would be 0-3 so you'd need to subtract 1 from the node-number in your expression for the check to work properly.

So... not a DataStage issue or question. Probably better to ask this on an Autosys support forum. Regardless, my previous reply still stands - there's obviously something incorrect in your script if it only works correctly when run manually.

A little lost. :? Is this actually a DataStage question or are you simply having a problem using Autosys to run a shell script? For the latter I'm going to give you some standard advice for shell scripts, especially for ones that 'work perfectly' when run manually but fail in some fashion when run u...

"node-number" would be 1,2,3 or 4 in your case. As for the Netezza specific questions you keep asking, those should be going to your DBA.

Interesting... trying to make up for some of the discrepancies between PX and Server handling, it would seem.

OK, so no. Your question implied that they were. In that case I would stick with the previous explanation.

qt_ky wrote:I really doubt that the IP address is embedded anywhere in the software. We have changed IP addresses on our servers without incident.

Yah my bad, went back and searched for some of the topics I thought I was remembering on the subject and it is the host names that are the key.