Archives of the TeradataForum

Message Posted: Sun, 03 Jul 2005 @ 16:53:04 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Your target PPI table is a SET table with no unique index, and the partitioning key is not included in the PI. So for each row inserted, duplicate row check must probe all 120+ partitions.

For better insert performance, change one of the factors noted above:

Define the table as MULTISET (bypasses duplicate row check)

Include the partitioning key in the PI (limits checking to one partition)

Define a USI (duplicate index key check eliminates need for duplicate data row check)

	<Prev	Next>		<<First	<Prev	Next>	Last>>