1) Turn of Logging, Direct load insert (Skip the conventional path i.e skip evaluation buffer).
2) Pre-Sort the files – Sorting the file in the order of primary key index speeds up the indexing processing.
3) Drop & Re-Index if possible
4) Some times at ETL layer the partition/Parallelism also helps it break the files into data same as partitions in Oracle
5) If using RBS can assign a BIG RBS during the processing.
6) If partition table see if partition exchange is possible if not then DO Parallel inserts in different partitions.
7) Pre-Allocate extents to the staging table
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment