DSXchange

Not peek, pivot.

Nothing other than the obvious - that you don't have a column in your hashed file with the name you used in the CREATE.INDEX statement. You sure it was spelled correctly, case and all?

Sure, but keep in mind the experience level. Adding a sequence job and coding a BASIC routine into the mix adds complexity, typically beyond what people have done before. Always a first time, I know, but this really can be solved rather simply inside the existing job with a hashed file and two refer...

Okay... any wagers on what the next suggestion would be? No?

Add -mode NORMAL to your dsjob command and see if the behaviour is different. And remove the -wait if what you really want is the -jobstatus option - one or the other, not both.

It doesn't want the full path to a hashed file, it wants the VOC name. Which means you either need to create an 'Account' based hashed file or, after creating a pathed one in a directory, create a VOC entry for it as well. I don't have it handy, but I'm sure that if you searched for your CREATE.INDE...

Sorry, it changes quite frequently.

As noted, why give a filename when asked for a directory name? Ditch the "emp_temp.txt" part.

For deletes, just reverse the roles and check again. A hashed file miss would be a deleted record then.

Ray, they did. Satheesh, your problem has been asked and answered quite a number of times here already, so many times that I just don't feel the need to spell it all out again. Search for some unique part of your error message to see what I mean. For example, I did an Exact search for ORA-01036 (das...

Phantom just means 'background process'. Is there more to the message? If the job aborts, do you get a 'From previous run...' log message when you Reset the job?

Create a hashed file with two hard-coded key records from the second file, one for the start and one for the end value. Look them up and constrain your first file's passage through the transformer based on those two values.

Umm... you did notice that the issue you responded to is pushing two years old, did you not?

Specifically the Field() function can be used to break up your pipe delimited string. Because of the double pipes, you'll need to adjust your 'field number' accordingly. For example, "incident" would be the third (not second) field. The 'sub fields' could then be broken further down using the same f...

OK. That would have been something worth mentioning in your original post, we wouldn't have wasted time talking about trying to 'fix' a Sybase issue. So... the error itself isn't the issue, it's the fact that when you use the ODBC stage it doesn't capture 'some of the errors'? Not really sure what t...

Hmmm... I was going to make a snarky comment about 'headless users' never having heard the term before, but since Kim didn't...

I can only find mention of the term but no definition, anyone care to share it?