Re: Should total_pages be calculated after partition pruning and constraint exclusion?
| От | Tom Lane |
|---|---|
| Тема | Re: Should total_pages be calculated after partition pruning and constraint exclusion? |
| Дата | |
| Msg-id | 3065.1526415058@sss.pgh.pa.us обсуждение |
| Ответ на | Should total_pages be calculated after partition pruning andconstraint exclusion? (David Rowley <david.rowley@2ndquadrant.com>) |
| Ответы |
Re: Should total_pages be calculated after partition pruning andconstraint exclusion?
|
| Список | pgsql-hackers |
David Rowley <david.rowley@2ndquadrant.com> writes:
> It seems quite strange to me that we calculate total_pages before any
> partition pruning and constraint exclusion is performed during
> set_base_rel_sizes(). Wouldn't it be better to wait until after that's
> done so we don't mistakenly count relations we're not going to scan?
> Delaying the calculation until after set_base_rel_sizes() means it's
> still done in time before it's needed in set_base_rel_pathlists().
I think that when this was first done, there was not a separation between
those passes over the rangetable, so that such a simple fix would have
been impossible.
> The attached patch implements the change.
Please add to next CF. It's a bit late to be doing such things in v11,
IMO.
regards, tom lane
В списке pgsql-hackers по дате отправления: