DSXchange

Double-check with your official support provider. You may need to upgrade... or perhaps use the Connector stage. Or both.

Can you post the complete, unedited error message or messages? There's usually a "why" that goes along with the "what"...

Good to know, should have put that in your original post. So... you want: read - rejected - discarded? For that I would just use the aggregator. For the last two rows, multiply the number by -1 before sending it to aggregation and then sum() them. What about the skipped value if it is non-zero? You ...

Sorry but we'll need more detail than that to provide a legitimate answer. What are your rules for computing this "difference"?

That's an example of what - your input data? If so, what would your desired output data look like?

<split to your own post>

Unless all you have is the Server version, the Enterprise version comes with the MKS Toolkit which brings 'UNIX' commands to Windows, so more than likely you do have grep... or something grep-like.

Help us out and post the details of your problem.

Talk to your DBA and resolve the issue that is leaving indexes in an UNUSABLE state. Typically loading duplicates into a unique index is the reason behind that. As long as they are "unusable" you won't be able to bulk load into the table.

That would violate the "without using unix command" part of the question.

Do an exact search here for "the record is too big to fit in a block".

There is no "default" you can set anywhere and yes, what controls all those window positions is stored in the registry. All that's supposed to happen is it remembers what state things were in when you exited and then should try to put you back there on the next launch. If that's not happen...

:!: When you have an issue, please start your own post. If needed, link back to something similar if you think it will add value. I've done this for you - split out your post and linked it back to the post you replied to. Now, I did have to guess on your particulars - Version 8 and a UNIX server. If...

If you want one row being turned into multiple rows after the lookup, then the setting you used allows that to happen. If you do not want that, then either craft your lookup so it only returns a single row so no warning is thrown, or live with the warning as it throws away all of the 'extra' rows fr...

Thanks... that was the clarification I was looking for.

You could also use the 'new' APT variable that turns off the stripping... but I don't recall exactly what it is called at the moment.

Here you go:

http://www-304.ibm.com/support/docview. ... wg21386517

Sorry but what kind of "transformations" could you possibly do with a BLOB?