DSXchange

That return code of 1 from grep simply means 'No match found' and won't just happen when the file is empty.

Just... change them.

What is your target database?

Because you appended it.

As noted, the job name requires the invocation id but the stage and link name do not. Correct your call.

Define 'only duplicate data'.

m Do not post the same question in multiple forums marked as different job types. You either have a Server specific question or a PX specific one, not both. You just waste our time responding to questions with either an inappropriate answer or to a 'problem' that has already been resolved, like thi...

If anyone had a clue, they would respond. This is the kind of question that needs to go back to the vendor, so contact your official support provider and get the answer from the horse's mouth.

So, just to clarify... when you run it from the primary job do you actually supply an Invocation ID? Type it in or pick it from the dropdown? The only difference between starting a run from the 'primary' or one of the existing invocation ids in the Director is what the Invocation ID defaults to in t...

Have you tried just reading it as is? What result do you get? That 'embedded quote' should resolve itself normally, I would think.

And then realize what a pain it might be to use.

How are you managing this currently? I would build a Sequence job and just feed this job in a loop. UserVariables, Start/End Loop.

Break the job into two pieces as noted, so each aspect of it can be 'tuned' - first the extract and then the load itself. Where is the actual bottleneck in your process?

Sure looks like one.

Interesting... you've said all along the date was in YYYY-MM-DD format and yet a mask of YYYY/MM/DD works for you? Ok.

Anyway, glad it is sorted out regardless.

What did you find when you searched for a common portion of the error message? For example - "Unable to save EditServ.impl" returned serveral hits for me.

You transfer the 'actual file' by taking the data read from the FTP Stage and writing it to disk using a Sequential File stage.

I'm lost on how using a varchar doesn't get you the "real time" it takes to load the data.

And you don't change the NLS_DATE format, you match it and use a varchar. That's the only mechanism for loading a date that does not require the use of the TO_DATE function.