DSXchange

It's an unusual design.
You control when rows are committed; if you specify 1000 rows per transaction, then a COMMIT is sent for every 1000 rows arriving in stage A-orac8i.
All 10000 rows are processed into the Oracle table before the output link of A-orac8i is opened and its SELECT statement executed. Therefore, the ORDER BY clause will be applied to the entire result set of the SELECT statement.
If you're doing this purely to sort the data in the sequential file, why not investigate using either an operating-system sort facility (such as sort on UNIX), or a third party sort utility such as CoSort or SyncSort? (See www.datastagexchange.com for the CoSort plug-in for DataStage.)


Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

quote:Originally posted by Ray.Wurlod
[If you're doing this purely to sort the data in the sequential file, why not investigate using either an operating-system sort facility (such as sort on UNIX), or a third party sort utility such as CoSort or SyncSort? (See www.datastagexchange.com for the CoSort plug-in for DataStage.)


Free Intel versions of the sort utility (as well as some other goodies like cat, sed and tr) are available from the net.

Mihai

_________________________
desk direct:+441908448571

There's a MICROSOFT product called SFU(services For Unix) that is almost full unix tools on 386. I think that MICROSOFT is going to integrate it in windows server 2003. Works very nice