DSXchange

You don't need one, hard-code it to the same value that you hard-coded your made up key to.

Where? Doing what?

Well... there is that. Figured I'd just stick with "you can't" and see what came next.

Which is something they allegedly understand.

There is a function that will give you the day number or name for a date and you could use it to check for Mon-Fri v. Sat-Sun and call those 'business days'. Or you could use an honest-to-goodness date dimension. I don't have documentation access at the moment, I think it is simply DateToString() wi...

Show us what you put in your uvodbc.config file.

You can't.

'Not able to read it' doesn't really tell us anything. We'd need details on exactly what it is you are doing and what isn't working about it before anyone could help.

Convert them to a valid timestamp format first.

OK... so just how accurate is your example? Do you need to sum every three rows or all rows to produce this "JKL" row? Something else entirely?

That's not any kind of a pivot - rows to columns or vice versa - just looks like aggregation, sums of column values. What have you tried?

As noted, parameters are attributes of a job, they are not stand-alone components. Parameter sets, however, are.

Interesting that it (grouping) actually is required in the Parallel aggregator, so not a workaround but the resolution for the PX side of the house. Been far too long that I've had 'hands on' with it to recall.

I asked because it matters, the Aggregator stage in each product while providing similar functionality are not close to being the same implementation-wise. I don't have any DataStage access and my PC with all of my documentation on it is dead at the moment, so my ability to give specifics is... some...

Just to double-check, you posted in the Server edition forum but marked your question as a Parallel edition question. Which camp does it fall in?