DSXchange

Phooey, I was kinda liking the idea of a massage queue... was wondering where I could sign up.

Totally agree, Ken, it should. When I was warehousing we always had one and it came in quite handy for things like you mentioned - all things date were answered there. However, recent efforts I've witnessed seemed to not even consider it. Damn kids. The OP only asked for the total number of days in ...

Or just mark it as a Reject link and make sure it is last in the link order.

Yes, it will work just like that. Give it a shot and you'll see.

Only if you propagate it. And as far as I know, that hasn't changed in 8.x.

Ooo... haven't been able to say this for a long time.

From a UNIX command line, type "man paste" and (hopefully) all will be revealed.

You said both paths, I picked one.

Stage / Properties / Folder pathname: C:\xml_data

Outputs / Properties / Wildcard: *.xml

It told you 'sqllder not found' where? Doing what? It makes a difference. Typically, everything Oracle needs is defined in your dsenv file, which includes the necessary path, shared library and environment variable settings. Are you able to / have been using the Oracle stages and just have a sqlldr ...

Especially if this behaviour is specific to you / your PC and everything works fine for anyone else.

Ok, it can't find your files because you've entered the values wrong. The Folder pathname should just be "C:\xml" and then your Wildcard pattern should be "*.xml".

Give that a shot.

Which one? They're not the same thing. For FTP there is a dedicated stage but it won't do SCP. For either one, you can script something up and call it before/after job or in a Sequence job.

Yes, of course. I didn't pony that up (or the Oracle solution) as I thought the OP asked for a code solution 'using the transformer stage' and didn't mention the 'O' word.

Time to mark the thread as something - Resolved or Worked Around, regardless.

Or just use a Routine Activity stage with the proper triggers.

In a word, yes. An 'always on' job will have issues flushing the sequential file to disk since the stage is never 'closed'. Ernie probably needs to come and expound on this and let us know what a better choice would be.

Why would you need to do this all separate like that, read the file 10 times, 10 links, etc? I don't see a need for anything other than a single Transformer feeding the Pivot stage. You have the State key column with 10 data columns coming in to the transformer and the record goes back out with 20 d...