Re: Read-only tables to avoid row visibility check

Поиск
Список
Период
Сортировка
Искать
От
Merlin Moncure
Тема
Re: Read-only tables to avoid row visibility check
Дата
Msg-id
CAHyXU0yNbOSDBeDhGJ_CVQ4v46KsVqfUq4wjiRH8QzD3yRPCWA@mail.gmail.com
Ответ на
Список
Дерево обсуждения
Read-only tables to avoid row visibility check Seamus Abshere <seamus@abshere.net>
Re: Read-only tables to avoid row visibility check Merlin Moncure <mmoncure@gmail.com>
Re: Read-only tables to avoid row visibility check Melvin Davidson <melvin6925@gmail.com>
Re: Read-only tables to avoid row visibility check Seamus Abshere <seamus@abshere.net>
Re: Read-only tables to avoid row visibility check "David G. Johnston" <david.g.johnston@gmail.com>
Re: Read-only tables to avoid row visibility check Seamus Abshere <seamus@abshere.net>
Re: Read-only tables to avoid row visibility check Tom Lane <tgl@sss.pgh.pa.us>
On Mon, Feb 22, 2016 at 2:35 PM, Seamus Abshere  wrote:
> hi,
>
> https://wiki.postgresql.org/wiki/ReadOnlyTables mentions the possibility
> of `ALTER TABLE table SET READ ONLY`.
>
> Would this mean that row visibility checks could be skipped and thus
> index-only scans much more common?

Personally I don't see how that buys you very much.  Right now you can
VACUUM the table which will update the visibility map, allowing index
only scans to be chosen.  Visibility checks are also already optimized
away by the database over time via hint bits.

I think the optimization you are looking for is to have the database
exploit the fact that when the table is created and/or loaded in a
single transaction, it marks everything visible and valid by default
and then wipes it all away should the insert fail.

merlin

В списке pgsql-general по дате отправления
От: Nicklas Aven
Дата:
От: Melvin Davidson
Дата:
FAQ