Re: index scan backward plan question

On Tue, 2006-03-21 at 16:58 -0500, Kevin Murphy wrote:
> I have a table for which PG 8.3 is guessing wrong about a plan when the
> result set gets large.

There is no such thing as "PG 8.3".

>    ->  Index Scan Backward using merged_weight_date_idx on merged
> (cost=0.00..31295593.98 rows=141839 width=229) (actual
> time=3.888..10380.783 rows=500 loops=1)

The mis-estimation of the result set of the index scan is clearly a
problem -- have you run ANALYZE recently? Try re-running ANALYZE and
retrying EXPLAIN ANALYZE.

Otherwise, please provide the queries that trigger the problem and the
relevant schema definitions.

-Neil