BUG #18930: PostgreSQL fails to handle INTERSECT operation involving Empty Table

The following bug has been logged on the website:

Bug reference: 18930
Logged by: jinhui lai
Email address: jinhuilai(at)email(dot)ncu(dot)edu(dot)cn
PostgreSQL version: 17.5
Operating system: ubuntu 22.04
Description:

Hi, PostgreSQL Developers,
Please consider such queries:
SELECT * FROM empty_table INTERSECT SELECT * FROM not_empty_table;
SELECT * FROM not_empty_table INTERSECT SELECT * FROM empty_table;
Obviously, the above queries always returns an empty set. I think it should
return an empty set quickly. However, it waste much time. PostgreSQL fails
to optimize queries containing INTERSECT operations when one branch includes
an empty table. This leads to unnecessary memory consumption and query
cancellation, even though the result should be deterministically empty.
I think this is a common case in actual production scenarios. It's important
to clarify that users might not intentionally perform INTERSECT operations
on empty tables. Rather, they may be unaware that a table is empty . For
example, when data has been deleted by another user or process. If
PostgreSQL can address this performance bug, it would significantly improve
query efficiency and save users valuable time in such cases.
Thank you for your valuable time, looking forward to your reply!
Best regard,
Jinhui Lai
You can reproduce the bug as follow steps:
psql -U postgres -c "CREATE DATABASE testdb;"
pgbench -U postgres -i -s 1000 testdb
psql -U postgres
\timing on
\c testdb;
testdb=# CREATE TABLE empty_table AS SELECT * FROM pgbench_accounts LIMIT 0;
SELECT 0
Time: 3.111 ms
testdb=# SELECT * FROM empty_table INTERSECT SELECT * FROM pgbench_accounts;
aid | bid | abalance | filler
-----+-----+----------+--------
(0 rows)
Time: 16236.845 ms (00:16.237) -- expect less time
testdb=# explain SELECT * FROM empty_table INTERSECT SELECT * FROM
pgbench_accounts;
QUERY PLAN
-------------------------------------------------------------------------------------------------
HashSetOp Intersect (cost=0.00..5139362.35 rows=200 width=356)
-> Append (cost=0.00..4139360.25 rows=100000210 width=356)
-> Subquery Scan on "*SELECT* 1" (cost=0.00..14.20 rows=210
width=356)
-> Seq Scan on empty_table (cost=0.00..12.10 rows=210
width=352)
-> Subquery Scan on "*SELECT* 2" (cost=0.00..3639345.00
rows=100000000 width=101)
-> Seq Scan on pgbench_accounts (cost=0.00..2639345.00
rows=100000000 width=97)
JIT:
Functions: 6
Options: Inlining true, Optimization true, Expressions true, Deforming
true
(9 rows)
Time: 1.446 ms
testdb=# SELECT * FROM pgbench_accounts INTERSECT SELECT * FROM empty_table;
aid | bid | abalance | filler
-----+-----+----------+--------
(0 rows)
Time: 16303.520 ms (00:16.304) -- expect less time
testdb=# explain SELECT * FROM pgbench_accounts INTERSECT SELECT * FROM
empty_table;
QUERY PLAN
-------------------------------------------------------------------------------------------------
HashSetOp Intersect (cost=0.00..5139362.35 rows=200 width=356)
-> Append (cost=0.00..4139360.25 rows=100000210 width=356)
-> Subquery Scan on "*SELECT* 2" (cost=0.00..14.20 rows=210
width=356)
-> Seq Scan on empty_table (cost=0.00..12.10 rows=210
width=352)
-> Subquery Scan on "*SELECT* 1" (cost=0.00..3639345.00
rows=100000000 width=101)
-> Seq Scan on pgbench_accounts (cost=0.00..2639345.00
rows=100000000 width=97)
JIT:
Functions: 6
Options: Inlining true, Optimization true, Expressions true, Deforming
true
(9 rows)
Time: 1.732 ms