Re: Introducing floating point cast into filter drastically changes row estimate

On Wed, Oct 24, 2012 at 10:06 AM, Merlin Moncure <mmoncure@gmail.com> wrote:
> I was chasing down a query that ran fine in 8.1 but had an near
> infinite runtime in 9.2.  It turned out to be from a bad filter
> estimate that is surprisingly simple to reproduce:

Testing some more it turns out that this isn't really a bug -- it's
just a general property of expression based filters: the planner
assumes they are very selective when in this case they are not.

merlin