Date: Mon, 20 May 2024 17:59:59 +0200
From: hubert depesz lubaczewski <depesz@depesz.com>
Sender: depesz@depesz.com
To: pgsql-bugs mailing list <pgsql-bugs@postgresql.org>
Subject: UNION removes almost all rows (not duplicates) - in fresh build of
 pg17!
Message-ID: <Zktzf926vslR35Fv@depesz.com>
Reply-To: depesz@depesz.com
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
Archived-At: 
 <https://www.postgresql.org/message-id/Zktzf926vslR35Fv%40depesz.com>
Precedence: bulk

So, test case is trivial:

#v+
$ select count(*) from pg_class where relkind = 'r';
 count
───────
    68
(1 row)

$ select count(*) from pg_class where relkind = 'i';
 count
───────
   164
(1 row)

$ select count(*) from ( select * from pg_class where relkind = 'r' union select * from pg_class where relkind = 'i' );
 count
───────
     1
(1 row)
#v-

explain shows unexpected:

#v+
$ explain (analyze) select count(*) from ( select * from pg_class where relkind = 'r' union select * from pg_class where relkind = 'i' );
                                                              QUERY PLAN
──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────
 Aggregate  (cost=42.43..42.44 rows=1 width=8) (actual time=0.108..0.109 rows=1 loops=1)
   ->  Unique  (cost=0.00..39.53 rows=232 width=236) (actual time=0.013..0.105 rows=1 loops=1)
         ->  Append  (cost=0.00..39.53 rows=232 width=236) (actual time=0.012..0.099 rows=232 loops=1)
               ->  Seq Scan on pg_class  (cost=0.00..19.19 rows=68 width=263) (actual time=0.012..0.058 rows=68 loops=1)
                     Filter: (relkind = 'r'::"char")
                     Rows Removed by Filter: 348
               ->  Seq Scan on pg_class pg_class_1  (cost=0.00..19.19 rows=164 width=263) (actual time=0.002..0.030 rows=164 loops=1)
                     Filter: (relkind = 'i'::"char")
                     Rows Removed by Filter: 252
 Planning Time: 0.161 ms
 Execution Time: 0.129 ms
(11 rows)
#v-

It seems that I get 232 correct rows from Append, but then Unique
removes all of them, except for one?

For whatever it's worth returned row is:

#v+
$ select * from pg_class where relkind = 'r' union select * from pg_class where relkind = 'i' \gx
─[ RECORD 1 ]───────┬───────────────────────
oid                 │ 2619
relname             │ pg_statistic
relnamespace        │ 11
reltype             │ 10029
reloftype           │ 0
relowner            │ 10
relam               │ 2
relfilenode         │ 2619
reltablespace       │ 0
relpages            │ 19
reltuples           │ 410
relallvisible       │ 19
reltoastrelid       │ 2840
relhasindex         │ t
relisshared         │ f
relpersistence      │ p
relkind             │ r
relnatts            │ 31
relchecks           │ 0
relhasrules         │ f
relhastriggers      │ f
relhassubclass      │ f
relrowsecurity      │ f
relforcerowsecurity │ f
relispopulated      │ t
relreplident        │ n
relispartition      │ f
relrewrite          │ 0
relfrozenxid        │ 730
relminmxid          │ 1
relacl              │ {pgdba=arwdDxtm/pgdba}
reloptions          │ [null]
relpartbound        │ [null]
#v-

As for test environment:

Debian testing on amd64, self-compiled from git HEAD at d2a04470aa6401c1938cc107e0b2c56c22a2321f

Did I do something wrong?

Best regards,

depesz