| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | David Rowley <dgrowleyml(at)gmail(dot)com> |
| Cc: | Richard Guo <guofenglinux(at)gmail(dot)com>, huyajun <hu_yajun(at)qq(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org, feichanghong(at)qq(dot)com |
| Subject: | Re: Potential Issue with Redundant Restriction Clauses in get_parameterized_baserel_size for PARTITIONED_REL |
| Date: | 2024-11-27 02:26:15 |
| Message-ID: | 1763158.1732674375@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
David Rowley <dgrowleyml(at)gmail(dot)com> writes:
> It would be good to understand why get_parameterized_baserel_size()
> bothers accounting for the baserestrictinfo quals and does not just do
> clauselist_selectivity() on param_clauses alone and multiply by
> rel->rows (which should already account for the baserestrictinfo).
One reason is that clauselist_selectivity does not necessarily
treat all the clauses independently. For example, if "x > 1"
is in one list and "x < 4" is in the other, you'll get very
different (and worse) results if you keep the lists separate
and just multiply their selectivities together.
We do have per-RestrictInfo selectivity caching that eliminates
most of the apparent inefficiency in this.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michail Nikolaev | 2024-11-27 02:27:00 | Re: Strange assertion in procarray.c |
| Previous Message | David Rowley | 2024-11-27 02:10:07 | Re: Potential Issue with Redundant Restriction Clauses in get_parameterized_baserel_size for PARTITIONED_REL |