DSXchange

Yes, 'clear' is a transactional delete of all rows in the table. You don't actually need to read an empty table, any method to send zero rows to the stage will work. For example, starting from a Transformer stage with the constraint set to @FALSE works great. All you need is a bogus stage variable t...

Writing to a hashed file could slow a job over time, reading from one won't. I'm sure this is just another manifestation of 'back pressure' as Ken calls it. In this case, probably from holding a large number of rollback segments over the course of the load. First suggestion for this job design would...

Specifically, reset the aborted job and then let us know if an entry labelled 'From previous run...' is logged. If it is, post the contents.

Post your syntax.

Pretty much any one, actually. What one, if any, does your company use? Leverage the 'dsjob' command line utility to allow it to run your jobs. This executable is quite thoroughly explained in either of the Developer's Guides in the Command Line Interface section near the end.

I doubt there is anything wrong with 'OCI' but how you are using it. The preferred way to load large numbers of records is through a bulk loader. If all you are doing is inserts, why not give that a shot? However, you never answered the question posed as to what your update action is, did you? Besid...

'Cannot access' how, exactly? I clicked on the link you posted and went right there. Try it again.

Substring comes to mind. Also Field (unless that's not supported in a PX job?):

Actually, you can but it's not elegant. Do one 'Next Saturday' with the proper parameters and then after it runs, remember to schedule in the next one. Lather, rinse, repeat. Either that or say (this month) the 'Next 6th' then 'Next 13th' then 'Next 20th' (etc) and repeat only once a month. Better b...

After making the dsenv changes, did you stop and restart DataStage?

As I said.

Curious how this is a 'workaround'. You asked if, we said no, case closed. Resolved.

And just out of curiousity, are there actually any namespace declarations in your incoming XML files?

Have you read the Reference Guide for the stage? Done things like add the two $APT_ORCHHOME variables where and how it shows? Best to let us know what you've done so people can advise you on what's still left.

And, out of morbid curiousity, why not use the Oracle Enterprise stage?

Interesting... I thought they meant the ODBC stage did not support the capture of returned result sets from a stored proc. I've never had a reason to look into this as it's not something I generally need to do or do for that matter. However, in reading the online Help it does say the ODBC stage supp...

Dual core will require two licenses - even hyper-threaded CPUs will. And just as an aside, now that we've gone crashing thru the Topic Barrier - Narsi, that cnet article is over a year old and the CPU wars flip-flop 'winning' sides quite a bit. Intel's new 'Core Duo' chips are the latest King of the...