Re: More stable query plans via more predictable column statistics

On Tue, Mar 15, 2016 at 4:47 PM, Shulgin, Oleksandr <
oleksandr(dot)shulgin(at)zalando(dot)de> wrote:

> On Wed, Mar 9, 2016 at 5:28 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>
>> "Shulgin, Oleksandr" <oleksandr(dot)shulgin(at)zalando(dot)de> writes:
>> > Yes, I now recall that my actual concern was that sample_cnt may
>> calculate
>> > to 0 due to the latest condition above, but that also implies track_cnt
>> ==
>> > 0, and then we have a for loop there which will not run at all due to
>> this,
>> > so I figured we can avoid calculating avgcount and running the loop
>> > altogether with that check. I'm not opposed to changing the condition
>> if
>> > that makes the code easier to understand (or dropping it altogether if
>> > calculating 0/0 is believed to be harmless anyway).
>>
>> Avoiding intentional zero divides is good. It might happen to work
>> conveniently on your machine, but I wouldn't think it's portable.
>>
>
> Tom,
>
> Thank you for volunteering to review this patch!
>
> Are you waiting on me to produce an updated version with more comments
> about NULL-handling in the distinct estimator, or do you have something
> cooking already?
>

I've just seen that this patch doesn't have a reviewer assigned anymore...

I would welcome any review. If we don't commit even the first part
(bugfix) now, is it going to be 9.7-only material?..

--
Alex