DSXchange

What do you mean by "and reverted back"?

Kryt0n wrote:that looks more like a field delimiter as | and record delimiter as that abc thing...

Hmmm, now that you mention it - true.

Still, the short answer is keep track of where you last comitted and skip records the next run until you get past that point. This assumes you are loading from a static source.

But as Ray noted, it should also be telling you "Invalid character conversion found converting XXX to ASCL_ASCII, substituting". Without that, we're completely punting in the dark.

That would have been nice to mention in the first post. Reverse the process then and 'post-process' the file. Forget the Filter option.

Hmmm... just mean to put the file through something that will convert those end delimiter strings to something more... normal. For example, the 'tr' utility in UNIX, or perhaps awk or perl. And this is exactly the kind of thing that the 'Filter' option in the Sequential File stage is for: read the f...

Where in the blue blazes does stiff like this come from?

Is there an "end delimiter string" property in the stage? If not, it seems to me you'd need to preprocess the file (perhaps via the filter) to replace those delimiter characters with something the stage can handle.

Why do you think you need to access this? What kind of information are you after, specifically? ps. It's a Universe database, not just "a hashfile".

For something like that, I'd see if my DBA could do something about it. A little chunk of PL/SQL would probably do the trick nicely, I would think.

chulett wrote:I could swear I've seen this before, and seem to think it is something Linux specific but can't recall the gory details for the life of me.

I remind me of me.

viewtopic.php?t=128498

It could simply mean the data / cursor / whatever is still cached on the Oracle side for the second run. Have your DBA lend a hand, trace the execution both times and see.

OK, just got a clarification from the developer that it may very well have been extended ASCII on the dev side as well. He has a habit of pulling UNIX files over to his PC to look at them and then pusing them back and I think that process may have cleaned up that character. He'll run it again and ch...

Just been brought an interesting problem that I can't seem to solve quite yet. We have a bog-standard 7.5.2 Server job writing output to a sequential file. We have dev and qa projects on the same server and engine install and in dev it behaves itself just fine. The problem comes when it's run after ...

Fair enough, had to ask. I wonder about trailing characters still... does String_Column[1,1]='A' work ok?

What data type is this field? A CHAR field, for example, would have trailing spaces so you'd have to add them to the criteria or trim them in the constraint.