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

>>>>> "Tom" == Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> writes:

Tom> Also, the way that the value is calculated in the
Tom> samples-not-all-distinct case corresponds to the way I have it in
Tom> the patch.

Ahh, gotcha. You're referring to this:

/*
* If we estimated the number of distinct values at more than 10% of
* the total row count (a very arbitrary limit), then assume that
* stadistinct should scale with the row count rather than be a fixed
* value.
*/
if (stats->stadistinct > 0.1 * totalrows)
stats->stadistinct = -(stats->stadistinct / totalrows);

where "totalrows" includes nulls obviously. So this expects negative
stadistinct to be scaled by the total table size, and the all-distinct
case should do the same.

Objection withdrawn.

--
Andrew (irc:RhodiumToad)