Re: Bogus ANALYZE results for an otherwise-unique column with many nulls

Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com> writes:
> On 5 August 2016 at 21:48, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> OK, thanks. What shall we do about Andreas' request to back-patch this?
>> I'm personally willing to do it, but there is the old bugaboo of "maybe
>> it will destabilize a plan that someone is happy with".

> My inclination would be to back-patch it because arguably it's a
> bug-fix -- at the very least the old behaviour didn't match the docs
> for stadistinct:

Yeah. I suspect that situations like this are pretty rare, or we'd have
recognized the problem sooner. And at least for Andreas, it'd be fixing
a real problem. I'll apply the back-patch unless I hear objections in
the next couple of hours.

> Additionally, I think that example is misleading because it's only
> really true if there are no null values in the column. Perhaps it
> would help to have a more explicit example to illustrate how nulls
> affect stadistinct, for example:

Good idea, will do.

regards, tom lane