| From: | Zsolt Parragi <zsolt(dot)parragi(at)percona(dot)com> |
|---|---|
| To: | Jeff Davis <pgsql(at)j-davis(dot)com> |
| Cc: | pgsql-hackers(at)lists(dot)postgresql(dot)org, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp> |
| Subject: | Re: Inconsistency with EXPLAIN ANALYZE CREATE MATERIALIZED VIEW |
| Date: | 2026-11-19 09:44:00 |
| Message-ID: | CAN4CZFPY2uHkSZYS_n72jZ1aR=29fv93N7P=54_gWCsFjGLsrQ@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hello
I ran into this while working on an event trigger issue in the same
code path [1], and only found this thread after writing a fix, so here
is a patch for option 2.
It follows the previously discussed approach: with ANALYZE, EXPLAIN
first creates the empty view with create_ctas_nodata(), in the
caller's context,. Then it plans and runs the query as a
security-restricted operation with a restricted search_path, so it
also covers functions executed by the planner. A new IntoClause field
tells the intorel receiver to fill the existing relation instead of
creating it. Like RefreshMatViewByOid(), it keeps the view open while
filling it, otherwise a function evaluated by the planner could, for
example, create an index on it that intorel_receive() doesn't
maintain.
[1]: https://postgr.es/m/CAN4CZFOycbakF5swg0i_hH2dZTkMP2%2BoA%3DnoU0VFFFe3687ojw%40mail.gmail.com
| Attachment | Content-Type | Size |
|---|---|---|
| v1-0001-Restrict-EXPLAIN-ANALYZE-CREATE-MATERIALIZED-VIEW.patch | application/octet-stream | 13.1 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Zsolt Parragi | 2026-11-19 09:51:55 | EXPLAIN ANALYZE CREATE TABLE AS EXECUTE runs non-SELECT prepared statements |
| Previous Message | Zsolt Parragi | 2026-11-19 09:37:35 | Fire event triggers for relations created by EXPLAIN ANALYZE (CTAS, SELECT INTO, MAT VIEW) |