Re: Performance Question on Joins

Both of your SQL statements are not likely to work well, 'cause your primary search path is a column of table B that is not the first in the index.

The first statement would probably be processed through a full scan on table B, followed by an index unique scan on table A; the second one would involve a full scan on table B and a range scan on the primary key index of table A: the point is the 100,000 rows that will have to be scanned in table B.

If you create an index on key_3, or if you put key_3 as the first field in table B's primary key index, both queries would improve significantly their response times.

Best regards Received on Wed Sep 13 1995 - 00:00:00 CEST