Datastage Sort in conjunction with Oracle Sort

[kool_cons]

We are using Oracle read parallesim in datastage to run the same query on multiple nodes..Later we have to sort the data to remove dups.Is it better perform if we do Oracle Sort and Again Datastage Sort or just do datastage Sort? If the data is sorted in oracle before it goes to Datastage Sort,Is datastage clever enough to know that data is already sorted and will not take more time to Sort entire data again in datastage?

Thanks In Advance.

[ArndW]

Sorting partially sorted data tends to be faster than sorting random data. You can insert a sort stage with "Don't sort, previously sorted" for the already sorted columns to let DS know that it doesn't need to sort data.