po 26. 6. 2023 v 19:48 odesílatel James Coleman <jtc...@gmail.com> napsal:
> On Mon, Jun 26, 2023 at 1:45 PM Pavel Stehule <pavel.steh...@gmail.com> > wrote: > > > > > > > > po 26. 6. 2023 v 19:43 odesílatel Pavel Stehule <pavel.steh...@gmail.com> > napsal: > >> > >> Hi > >> > >> po 26. 6. 2023 v 19:41 odesílatel James Coleman <jtc...@gmail.com> > napsal: > >>> > >>> Hello, > >>> > >>> Have we ever discussed running an analyze immediately after creating a > table? > >>> > >>> Consider the following: > >>> > >>> create table stats(i int, t text not null); > >>> explain select * from stats; > >>> Seq Scan on stats (cost=0.00..22.70 rows=1270 width=36 > >>> analyze stats; > >>> explain select * from stats; > >>> Seq Scan on stats (cost=0.00..0.00 rows=1 width=36) > >>> > >>> Combined with rapidly increasing error margin on row estimates when > >>> adding joins means that a query joining to a bunch of empty tables > >>> when a database first starts up can result in some pretty wild plan > >>> costs. > >>> > >>> This feels like a simple idea to me, and so I assume people have > >>> considered it before. If so, I'd like to understand why the conclusion > >>> was not to do it, or, alternatively if it's a lack of tuits. > >> > >> > >> I like this. On the second hand, described behaviour is designed for > ensuring of back compatibility. > > > > > > if you break this back compatibility, then the immediate ANALYZE is not > necessary > > I don't follow what backwards compatibility you're referencing. Could > you expand on that? > Originally, until the table had minimally one row, the PostgreSQL calculated with 10 pages. It was fixed (changed) in PostgreSQL 14. https://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=3d351d916b20534f973eda760cde17d96545d4c4 Regards Pavel > Regards, > James Coleman >
