DSXchange

The preference would be to use the orchdbutil command to import the table metadata as that is the most accurate way for PX.

Nowhere in DataStage, that would need to be set by a UNIX Admin.

Still not showing up clearly. Code tags, peoples, code tags maintain whitespace. And use Preview to check before posting. table A | | | table B ---------lookup stage -______matched record(file1) | | | Unmatched records (file2)

If you read the file as one long string field, can you recogize each type of record easily? Say by checking the first few bytes for 'HDR' or 'FTR' (etc).

Put your expression in a 'test harness' (a small routine) then test to ensure you are getting what you think you are getting from it:

There is no 'date format' in an Oracle database, per se, unless you are attempting to match the NLS_DATE format. Please post the sql and clarify how the DATE column is defined in the job itself.

Why did you feel the need to change the driver name? That surely is the path to the Dark Side...

Ok, good. Saves me from going into my lecture on how Oracle dates are not stored in any format and what they look like when selected is up to you or your tool of choice.

My example showed that. "MON" = upper-case three character month, "mon" = lower-case. Pick whichever you like.

Search the forums for either 'ETLStats' or 'DSGetLog', both should reveal the bajillion conversations we've had on this subject.

Ray? Is that you?

The post you linked to explains what to do when your DataStage server is on Windows. For UNIX? Different story - you use the [Text] driver as far as I know. Make a copy of it to a new name and try that.

Would need more information on how you are using the date, but basically you can substring the 'yymmdd' portion out of the filename, then use functions to convert it, if need be.

Please don't start a new post for something you've already asked about:

http://www.dsxchange.com/viewtopic.php?p=297394

You should have just asked for a clarification there.