DSXchange

I think the END statements are complete. Three conditions, three END statements. But you are right about checking the first condition twice, no need for that, and hence your code takes over, as always. Ray ji, you rule

You are not binding the variables correctly. We need to look at your query and specially your where predicate.

Craig, you beat me to it. He is write. You need a nested if then else. If ((Arg1 <> Arg2) And (Index(Arg1, 'NETWORKING_EXCHANGE',1) >= 1)) Then Ans ='SUB' End ELSE If ((Arg1 <> Arg2) And (Index(Arg1, 'NETWORKING_EXCHANGE',1) = 0)) Then ...

It should work in a stage variable as well. But it depends upon how many stage variables you are using, the order you define the stage variables etc etc.
Are you still running in parallel though? It wasnt a parallel vs. sequential issue was it?

As for the why and when, refer here.

Very funny ganesh123. Thats not called re-invention. Thats called enhancement.

Really, thats strange. Where are you giving the derivation, in a stage variable???
Run in sequential mode just for testing purposes and see if the result persists.

No routine is required in this case. This scenario can be handled perfectly and Krazykoolrohit gave a demo on that.
If your absolutely have to use a routine, then go for narasimha's suggestion. No need to re-invent the wheel.
My 2 centy!

It sets the value to null by using SetNull() if nothing is found. IF you want something else then you need to replace SetNull() in my code with whatever you want to populate.

Thats why I specified in my previous post that it can get him started, but should not be considered as the final solution.

See if this works for you: If (INDEX(in.Col, "L", 1) > 0) THEN Field(in.Col," ", Count(in.Col[1,INDEX(in.Col,"L",1)]," "),2)[1,INDEX(in.Col,"L",1)] ELSE IF (INDEX(in.Col, "cc", 1) ...

If data is like 12345$ABC6789 then populate data as 12345A6789 if data is like 12$BCD3456789 the populate data as 12B3456789 if data is like 1234567$CDE89 the populate data as 1234567C89. Look at your final data. It contains the first alphabet right after the $ sign. Thats what the substring is for.

I did not get that. You mean you want to put the filter in the sql select itself? If yes then get the max id of your target table and pass that as a job parameter to the where predicate of the sql select.

THere is a document that Craig provided in a post over at developersnet. You can request Craig to send the document to you.

I was missing a 'then' after the if statement. I fixed my original code.