psql's FETCH_COUNT (cursor) is not being respected for CTEs
Jakub Wartak <jakub.wartak@enterprisedb.com>
From: Jakub Wartak <jakub.wartak@enterprisedb.com>
To: PostgreSQL Hackers <pgsql-hackers@postgresql.org>
Date: 2023-01-04T12:10:20Z
Lists: pgsql-hackers
Attachments
- 0001-psql-allow-CTE-queries-to-be-executed-also-using-cur.patch (application/octet-stream) patch 0001
Hi -hackers,
I've spent some time fighting against "out of memory" errors coming
out of psql when trying to use the cursor via FETCH_COUNT. It might be
a not so well known fact (?) that CTEs are not executed with cursor
when asked to do so, but instead silently executed with potential huge
memory allocation going on. Patch is attached. My one doubt is that
not every statement starting with "WITH" is WITH(..) SELECT of course.
Demo (one might also get the "out of memory for query result"):
postgres@hive:~$ psql -Ant --variable='FETCH_COUNT=100' -c "WITH data
AS (SELECT generate_series(1, 20000000) as Total) select repeat('a',
100) || data.Total || repeat('b', 800) as total_pat from data;"
Killed
postgres@hive:~$ tail -4 /var/log/postgresql/postgresql-14-main.log
[..]
2023-01-04 12:46:20.193 CET [32936] postgres@postgres LOG: could not
send data to client: Broken pipe
[..]
2023-01-04 12:46:20.195 CET [32936] postgres@postgres FATAL:
connection to client lost
With the patch:
postgres@hive:~$ /tmp/psql16-with-patch -Ant
--variable='FETCH_COUNT=100' -c "WITH data AS (SELECT
generate_series(1, 20000000) as Total) select repeat('a', 100) ||
data.Total || repeat('b', 800) as total_pat from data;" | wc -l
20000000
postgres@hive:~$
Regards,
-Jakub Wartak.
Commits
-
Further review for re-implementation of psql's FETCH_COUNT feature.
- c21d4c416ad6 17.0 landed
-
Re-implement psql's FETCH_COUNT feature atop libpq's chunked mode.
- 90f5178211cd 17.0 landed
-
Support retrieval of results in chunks with libpq.
- 4643a2b265e9 17.0 landed
-
Attempt to fix newly added Memoize regression test
- a3a836fb5e51 17.0 cited