Re: Proposal - improve eqsel estimates by including histogram bucket numdistinct statistics

"Nathan Boley" <npboley@gmail.com> writes:
> Given that, wouldn't it be smarter to consider a value as an mcv
> candidate iff it has a density greater than 0.005, rather than having
> a count greater than 1.5*average?

Yeah, perhaps ... want to experiment with that?  Though I'd be a bit
worried about how to get the threshold right, seeing that it will
depend a whole lot on what the user has selected for random_page_cost
and other parameters.
        regards, tom lane