On 9/23/25 12:43, Andrei Lepikhov wrote:
But is it the same for the 'distinct' statistics? It seems you should love it - the number of groups in GROUP-BY, DISTINCT, and even HashJoin should be estimated more precisely, no?
I think it has more potential, and I would love to use this weapon, but I haven't had the opportunity yet. It would be interesting to know how much it is used in real life.
To get back to the topic of partitioned statistics, do you know if SQL Server is smart enough to handle this case [1] that we discussed last year? (with filtered statistics)
[1] https://www.postgresql.org/message-id/flat/b860c71a-7cab-4d88-ad87-8c1f2eea9ae8%40dalibo.com
