DSXchange

Connecting to the new version project created the hashed files and VOC entries, making it safe to simply move over the 'old' hashed files afterwards.

While it may be similar, it sure isn't the same. Please start your own post on this topic so we know all of your particulars and the actual error(s) you are seeing. The more gory details you can provide, the better. In your new post. Not here.

It's probably something that should have happened in the early version that they "fixed" in the later release. That "informational message" is what you now see - the number of rows affected. There may be some setting in the DB or ODBC options to turn that off, be curious what sup...

64bit operating systems fully support 32bit applications. The other way 'round won't work but you'll be fine.

Glad you found this too, but less glad that you decided to hijack the thread you found it in. Please start your own post on this subject.

Use two cursors and ensure the inner one gets properly released each time.

Why not copy it locally and then run the setup? Many things won't install properly from a 'network location'.

Search, I needed to do the same thing many years ago and posted exactly how to do that.

ps. Ray was joking... and perhaps meant Robert Rankin rather than DA.

Not a TX nor SQL Server person, but if you google for that OLE DB error it looks to be a unique index / primary key duplicate value issue.

What do you need this "export" for? You may have the istool utility, depending on your exact 8.x release, but that may not help... depending on what you intend to do with this "export of the job in XML format".

Odd... glad you got it sorted out, or at least working again... for now.

And are we talking a Server routine or a Parallel one as the post is marked?

No idea. You are the victim of a "communication error" as noted in the log messages you posted, hence the suggestion to involve the first line of defense for issues of that nature. And I think your DB2 DBA is leading you astray, not sure where they got all of the errors they reported back ...

So... you don't actually need to "transform" anything but rather... "detect" what characterset is in use, row by row, in a particular column?