DSXchange

I would guess some kind of permissions issue, something that was preventing the executables from being regenerated by the compiler, so you were technically still running the old version.

Seeing as how there are still a 'Universe' repository and file permissions in play in the 8.x world, I would certainly think so. Can't say if that is still the complete answer, however.

Probably would have been better to have asked this over in the FAQ posting itself rather than here.

Patch number, please.

You do an "outer join" by not checking the status of the lookup, but that's not a full outer join. For that you'd need to flip the streams and run the data through that way as well, then merge the results.

What exactly happens right now when "there are no rows that can be deleted or inserted"? What causes the abort?

Use a hex editor and let us know what these "square values" actually are.

The answer to the first question is 'yes'. For the second, not sure what you mean by "to do so? You can't abort on zero rows as nothing would be triggered then. Check 'after job' instead.

I've had and solved that same error. Search for "Object reference not set to an instance of an object" and see if my fix works for you as well.

No, you should be able to create a new post right away as long as you're registered and not in as a 'guest'. If you still need help, post over in the Site/Forums forum and we'll see what we can do. And it's ok to hijack one over there as long as you explain why you needed to.

And just to address the last question, there is no such thing as space "in a folder", for Windows the space is managed at the drive letter level and is available to any folder on that drive letter.

Both the cause and solution to this problem have been posted here before - I know because I posted one of them. That part wouldn't have changed in 8.x I don't believe.

Please search.

Then it should work fine as long as all jobs have the write stage set to append.

It was when you replied, Arnd, this is a completely unrelated hijack by a new poster... but at least it's in the right forum.

Mutiple processes cannot write to the same sequential file at the same time, what you're doing will only work if the jobs run in a serial fashion - do they?

Was the sequence originally set to Validate Only and you changed it? If so, are you positive it now says Run and that you've compiled the sequence job? And that you're running the same one you are looking at?