DSXchange

Raji - were you able to solve this? Still need help with it?

OK, that's a conversation for another post.

Did you try adding a Transformer?

That would have been excellent information to lead off with rather than make us dig for it.

Assign a sequential number to each record as it is read, that could be as simple as using @INROWNUM. Then join on this generated 'key' field.

So, one of three questions answered. How about the last one - without more information from you, we can't really help I'm afraid.

OK... what is the relationship between the two files? Do they share any keys in common? If not, can you state in words how they should be 'joined'?

Yes, that would be the full service solution. That would allow you to check for differences and/or 'missing' records from either side via null checks as Ray notes.

I am using Aqua to view the oracle tables. Hmmm... Aqua Data Studio? Rather than any kind of grid-based viewer like that or Toad, what happens if you simply use sqlplus rather than that tool? What value do you get then? Just trying to ensure you are solving a real problem rather than a perceived pr...

When we are running the job using dsjob command it is failing with error "PIN2 not initialized". Question - does this imply that you do not get the error if you run the job from the Director? :? Regardless - "IDENT1" is a transformer that the compiler added to your job between y...

Do you need to handle records missing from one file or the other? Meaning they exist in File1 but not File2 or vice versa...

What have you tried so far? What issues are you having? On the surface, it's just a pipe-delimited file which shouldn't really provide any challenges in the reading dept. I'm assuming that first record is the header row and can be skipped. If there's more to this than meets the eye, please clarify f...

You've had lots of help so far... what is your issue at this point? What remains to be solved?

First one in or out? @INROWNUM=1 or @OUTROWNUM=1 can work but of course you'll need to worry about partitioning when running on multiple nodes.

Q.E.D.