Re: Planner choosing NestedLoop, although it is slower...

On 07/12/2011 10:04 PM, Tom Lane wrote:
> Mario Splivalo<mario.splivalo@megafon.hr>  writes:
>>    Limit  (cost=0.00..415.91 rows=21 width=8) (actual
>> time=11263.089..11263.089 rows=0 loops=1)
>>      ->   Nested Loop  (cost=0.00..186249.55 rows=9404 width=8) (actual
>> time=11263.087..11263.087 rows=0 loops=1)
>
>> Why is planner using NestedLoops,
>
> Because it thinks the LIMIT will kick in and end the query when the join
> is only 21/9404ths (ie, a fraction of a percent) complete.  A NestLoop
> results in saving a lot of work in that situation, whereas hash-and-sort
> has to do the whole join despite the LIMIT.
>
> What you need to look into is why the estimated join size is 9400 rows
> when the actual join size is zero.  Are both tables ANALYZEd?  Are you
> intentionally selecting rows that have no join partners?

Hi, Tom.

Yes, both tables have been ANALYZEd. What do you mean, intentilnaly
selecting rows taht have no join partners?

    Mario