DSXchange

First, off to the Server forum for you! Second, can you detail out your job design for us, please?

... and at a moments notice it could change again... and again...

They are tables that hold tables - basically sub-tables / multivalued data. For the 'efficient' part, I would think we'd need to know the volumes involved.

Right, so a Type 2 to a Type 30. And yes, it seems like you could follow the 'option 3' steps.

For those of us without access to 9.1, what exactly are the 'differences' of which you speak? Nothing core to Server jobs changed in... forever. If the internal structure of the hashed file has changed, however, should be pretty simple to write a converter. I'd just be shocked if they did such a thi...

Got distracted and forgot to note that. D'oh.

Correct.

Right. Use explicit transformations, don't just take strings and shove them into other datatypes.

Why don't you share with us the details of your job design specific to this field - stages, datatypes, transformations, etc - then perhaps we can spot where things are going wrong.

I'm not sure any of this affects 'performance' per se. What is your actual Oracle datatype?

Your IsValid derivation is incorrect. Validate the string and only convert to a timestamp if it passes validation. As to the warning, show us your chopping.

You'd need some kind of equivalent to what say, Oracle does with the TIMESTAMP WITH TIME ZONE or TIMESTAMP WITH LOCAL TIME ZONE datatypes.

Did you ever get this resolved? Stumbled back on this today and just wanted to add that you could also use the Filter option of the Sequential File stage to feed in just the filenames using something like an "ls -1 <pattern>" O/S command.

Without it being "fixed width delimited" (i.e. fixed width fields but with a record delimiter) then they must ensure that those trailing spaces are preserved as the file is read byte by byte based on the sizes you define. Have the source system correct the file. You're certain about the re...