Re: Paralellizing Pl/sql inserts
Date: Wed, 23 Jan 2008 09:22:53 -0500
Lots of good suggestions here.
How is your big table partitioned? If it's something like 128 hash partitions it's going to be a pain, but if it's something more manageable, you could try splitting your 65 million row inserts into data by partition and do insert /*+ append */ bigtable (partition). Each of those can go in parallel (seperate stmts). Drop indexes first and recreate using massive parallel hint after inserts are done.
Let us know how it works out for you. It's an interesting problem.
FinnReceived on Wed Jan 23 2008 - 08:22:53 CST