Douglas Alan escribió:
> Hey, while I have you on the line, might you be so kind as to explain why
> this query is so slow? Shouldn't it just fetch the first row in the table?
> What could be faster than that?
>
> explain analyze select * from maindb_astobject limit 1;
> >
> > QUERY
> > PLAN
> >
> >
------------------------------------------------------------------------------------------------------------------------------------
> > Limit (cost=0.00..0.04 rows=1 width=78) (actual time=8091.962..8091.965
> > rows=1 loops=1)
> > -> Seq Scan on maindb_astobject (cost=0.00..3358190.12 rows=75426912
> > width=78) (actual time=8091.955..8091.955 rows=1 loops=1)
> > Total runtime: 8092.040 ms
> > (3 rows)
Maybe there's a bunch of pages with only dead tuples at the start of the
table? Maybe a lot of empty pages at the start of the table (If this
is 8.3 you have to consider sync_seqscan as well)
--
Alvaro Herrera http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.