DSXchange

Not really sure how I could elaborate much on it. Build a hashed file with all of the invoices from B. Stream in invoices from A and do a lookup to the B hashed file, when your lookup fails you need to delete the A invoice. I'd probably do the actual deletion in a separate step by landing the keys t...

Forget the specific numbers. Do an exact search for "Nonnumeric data when numeric required", you'll find over 50 other discussions of this issue. Check those out.

What did your official support provider suggest?

So Chuck, are you saying that's working for you? Sorry, but I don't recall ever seeing the original post until you replied to it today but more than likely I really wouldn't have had anything to add to the conversation other than moral support.

The target is what - another sequential file, database table? And the signs are leading or trailing signs? Also, are you sure they are all "missing" or is it just the plus signs?

Didn't you get other messages with this one, one like this for instance?

The current soft limit on the data segment (heap) size (XXX) is less than the hard limit (YYY), consider increasing the heap size limit

They've already done that. Aren't the "hard" values those configured in the kernel? There's different ways to find those, it depends on what flavor of UNIX you are running.

You'll find it doesn't ship with 8.1 but instead was delivered in a major patch/update for it, so you'd need to ensure you have that in hand and installed.

How are you currently getting jobs from dev to prod? Why not just reverse that? Or just go brute force and export everything from prod and import back into dev. Make sure you've got plenty of backups first!

You're not finding that because they are client tools, not server tools. If you have the latest release of 8.1 then you can use something called "istool" on the server from what I've read here.

How exactly did you "migrated our datastage server from 10.14.107.180 to 10.13.41.118"?

But... but... they'd never be able to use "DWTest" as a project name ever again!

[/kidding]

Start your own thread and provide the relevant details of your problem there, in particular exactly what "does not work!!" means.

dhruv12 wrote:If there is a way where this update/insert can take place in a dataset, then this would eliminate the need for the job which removes the duplicates.

Short answer - it cannot. It is sequential media, a data set not a data base.

Have you verified that that port is accessable? Meaning not blocked by something like a firewall? An exact search for "open ports" would probably help as well.