Re: Self contradictory examining on rel's baserestrictinfo

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Mon, Nov 25, 2024 at 4:39 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>> ... I think Robert
>> must have had constraint_exclusion = on for his examples a
>> couple messages back.)

> I definitely didn't change the value of that setting. It's possible I
> was running on an older branch, but I don't think it was anything
> ancient.

Hmm, constraint_exclusion has defaulted to "partition" for many
years now. But when I tried to reproduce your examples at [1]:

regression=# create table foo (a int);
CREATE TABLE
regression=# explain select * from foo where a < 1 and a > 1;
QUERY PLAN
-----------------------------------------------------
Seq Scan on foo (cost=0.00..48.25 rows=13 width=4)
Filter: ((a < 1) AND (a > 1))
(2 rows)

regression=# show constraint_exclusion;
constraint_exclusion
----------------------
partition
(1 row)

regression=# set constraint_exclusion = on;
SET
regression=# explain select * from foo where a < 1 and a > 1;
QUERY PLAN
------------------------------------------
Result (cost=0.00..0.00 rows=0 width=0)
One-Time Filter: false
(2 rows)

and similarly for some of the other examples (I didn't try
every one). Maybe your test table was partitioned??

regards, tom lane

[1] https://www.postgresql.org/message-id/CA%2BTgmoZqiCwHbZczXXLjucfuHi%3D7EahSyzEj5yrqYKMQ0QOL9Q%40mail.gmail.com