Re: Postgresql Materialized views
От | Merlin Moncure |
---|---|
Тема | Re: Postgresql Materialized views |
Дата | |
Msg-id | b42b73150801130907h49f3825esd21889e06cf0b3bc@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: Postgresql Materialized views (Simon Riggs <simon@2ndquadrant.com>) |
Ответы |
Re: Postgresql Materialized views
Re: Postgresql Materialized views |
Список | pgsql-hackers |
On Jan 12, 2008 4:19 PM, Simon Riggs <simon@2ndquadrant.com> wrote: > On Sat, 2008-01-12 at 13:27 +0100, Jean-Michel Pouré wrote: > > > Please pick-up this important issue for developpers. There is no need to > > concentrate on complex issues, when handling materialized views could > > boost somme web apps. by a factor of 10 or more. > > It's more complex than you think, but the main reason was that HOT was a > prerequisite for making summary tables work efficiently, which is only > now just about to go live into 8.3 +1 If you know how to write triggers, materialization techniques aren't all that difficult. The real technical limitation was not lack of materialization techniques (write triggers), but was dealing with the mvcc penalty. Previously to HOT, for summary tables I would redirect the trigger to insert to a 'roll up' table and move the data to the summary on cron or from an application event. Materialized views are syntax sugar (but still very sweet). merlin
В списке pgsql-hackers по дате отправления: