For Data Warehousing applications, where you tend to join very large tables,
hash joins can really speed things up. Even when joining domain tables to
large fact tables, nested loops can really bog things down.
There are other tuning methods you can employ, such as loading a small
domain table into memory (say, a nested table or index-by array) and then
use the memory object in the qyery instead of the actual table. I have seen
this technique speed queries up 800%.
- Dan Clamage
Received on Sat Feb 22 2003 - 19:51:41 CST
