DSXchange

Then I would think a Timestamp should be fine and there's no need for the scale you are using, which is causing your metadata mismatch. Try Timestamp(19) rather than Timestamp(23,3) and see if you can run it all the way through the job like that rather than starting off as a Char.

As already noted, it's not relevant.

You need to check the INTERIMJOBSTATUS first to determine after job if the job is on its way to aborting. And to connect to the current job, you don't need a handle but rather use DSJ.ME as the handle instead.

In other words, it seems to be a custom routine. Can you verify that?

Exactly. You use TO_DATE() to put data into a DATE field and TO_CHAR() to pull it back out again.

You need to answer Ray's question: what stage are you using to write to the hashed file? Something other than a Hashed File stage? I've never seen any error like that working with hashed files, so it makes me wonder too how exactly you are doing things. :? At the very least, explain what the heck &q...

Unfortunately, that's a PX suggestion and so not relevant here. Wade, can you post an example of what this "datetime" data looks like when you select it from SQL Server? I've not really worked with that database all that much. I'd also be curious what happens if you extract it defined as a...

Serious answer: avoid doing things that generate warnings. There's no reason a Server job (unlike a PX job) should generate any kind of warnings when it runs. I've had clients run over a thousand Server jobs every night to process their data without a peep out of any of them. Now, if you've got some...

Not true any more as of when? With which version?

New question = new post, please. This has nothing to do with "XML Single Line".

Do let us know how you end up working this out, please.

It has, and quite recently as well. A search for "vertical pivot" should turn them up. You'll even find one about the "Flush Record" I spoke of needing.

First suggestion - always use real examples when you illustrate your problem. It is troubling when someone posts something "like" what they need and you answer based on that example only to find out it really isn't what they need to do. So... you are using a sql profile rather than tuning ...

Afraid not, not until that "direction" is added in a future release as the pivot stage just does columns to rows. This needs a "stage variable concatenation" approach in a transformer to solve, probably with a dummy/trailer row to push out the last group when it hits EOD.

No, you want the "literal", a bind variable implies having something to bind to that value from every record so just use the job parameter there. Well... let's clarify that. You would use the Job Parameter in the SQL if you are sourcing from your database and want to constrain the source S...