Archives of the TeradataForum

Message Posted: Mon, 26 Sep 2005 @ 17:22:47 GMT

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

Madhukar,

This seems like very odd behavior. Do you have statistics on the large table's Loc_ID column? If not, I'd recommend collecting them.

Otherwise, you have 2 other possible solutions:

1) Assuming adequate data distribution and load factors, try creating a Single Table Join Index on the Large Table with a PI on Loc_ID

or

2) Move the 2600 rows out of the small table, collect stats while the table is empty, and move the 2600 rows back into the small table. This ought to result in the small table being duplicated. This solution is only recommended as a last resort. If the small table is used in many other queries, this will probably prove to be a bad solution to your problem.

In either case, you ought to open an incident with NCR to figure out why the optimizer is redistributing the large table.

Good luck,

Clay

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