On Fri, 2007-09-21 at 12:08 -0400, Tom Lane wrote:
> Csaba Nagy <nagy@ecircle-ag.com> writes:
> > Looking at Plan 2, it looks like the "limit" step is estimating wrongly
> > it's cost.
>
> The reason you get a bad plan is that this rowcount estimate is so far
> off:
That's true, but its not relevant, since the query would still be fast
even if that estimate was exactly right. With LIMIT 10, it wouldn't
matter how many rows were there as long as there were more than 10. The
true execution cost is limited, the cost model is not.
--
Simon Riggs
2ndQuadrant http://www.2ndQuadrant.com
