Parallel Extender - User Defined SQL in Parallel Job

seell · Post by **seell** » Mon Dec 13, 2004 9:58 pm

hi
I was used to DataStage V6 but now introduced to DataStage V7.0 - parallel extender as we require speed during data loading.

Noticed that there are 2 types of jobs : Server and Parallel; while the server job seems to operate in the same manner as in V6, the parallel version seems different.

The manual recommends to use Write method = load' when loading data to Oracle. What if I need to use 'user defined SQL' ?? In my case, one of the target fields is an oracle sequence.

Thank you

battaliou · Post by **battaliou** » Tue Dec 14, 2004 5:11 am

We had a similar issue loading data to Oracle in parallel. This is best done outside of datastage. If you are planning to use a sequence, you will have to collect your data into a single sequential text file else the sequence number generator can get quite tricky.

Remember you will need to generate a feeder id. This is the fastest method of loading data into Oracle.

Someswar Barkataki · Post by **Someswar Barkataki** » Wed Dec 15, 2004 9:59 am

Instead of "Load", select Write Method as "Upsert" and choose User-defined Update & Insert as the upsert mode. Now edit the Insert query as per your requirement.

Hope this helps,
Thanks,
Somesh