Horton

Re: unstable query plan on pg 16,17,18

Andrei Lepikhov <lepihov@gmail.com>

View thread View on postgr.es
From: Andrei Lepikhov <lepihov@gmail.com>
To: Attila Soki <atiware@gmx.net>, Laurenz Albe <laurenz.albe@cybertec.at>
Cc: pgsql-performance@postgresql.org
Date: 2026-02-27T15:35:31Z
Lists: pgsql-performance 
On 27/2/26 16:00, Attila Soki wrote:
> On 27 Feb 2026, at 09:15, Andrei Lepikhov <lepihov@gmail.com> wrote:
> I dont know the inner workings of analyze, is that normal that executing 
> analyze on unchanged data can flip the plan? Does analyze select a 
> random set of rows?

Yes, this is completely normal because ANALYZE works in a random way.
You can try using the Join-Order-Benchmark to see how running the 
ANALYZE command can change the execution time of the same query, 
sometimes by as much as ten times.

-- 
regards, Andrei Lepikhov,
pgEdge