Re: Index scan startup time

On Thu, Mar 30, 2006 at 02:51:47PM +0200, Steinar H. Gunderson wrote:
>On Thu, Mar 30, 2006 at 07:42:53AM -0500, Michael Stone wrote:
>> Yes. And the estimates are bad enough (orders of magnitude) that I can't
>> help but wonder whether pg could come up with a better plan with better
>> statistics:
>>
>>>>>>  ->  Index Scan using activity_pk on activity  (cost=0.00..40717259.91
>>>>>>  rows=6538650 width=8) (actual time=207356.050..207356.722 rows=100
>>>>>>  loops=1) Filter: ((state = 10000) OR (state = 10001))
>
>Isn't the rows=100 here because of the LIMIT?

Yes. I was looking at the other side; I thought pg could estimate how
much work it would have to do to hit the limit, but double-checking it
looks like it can't.

Mike Stone