Upsert without a Unique Constraint on DB2 Table

[shawn_2008]

Hello,
I have a Parallel job that reads from a sequential file and does an "Upsert" on a DB2 table based on user-defined sql. The upsert is done on a column which is not part of the primary key. Because of the way the table is partitioned, creating a unique constraint (index) on this column is not an option. When I execute the job, it inserts duplicate rows ( as far as this non-key column is concerned).

Is there a way to do an "UPSERT" without having a unique constraint on the table ?

Table: ACC_BOOKS
acc_book_no --> surrogate key
acc_book_isbn ---> ( want to do upsert based on this, cannot create unique constraint ont this)
acc_ .... --> other columns


Sequential File:
"ISBN","ACC_TRAN_COUNT"

Appreciate any help !

Cheers,
Shawn

[ray.wurlod]

Only by checking for existence of the value within your DataStage job.