Re: performance of insert/delete/update
От
Tim Gardner
Тема
Re: performance of insert/delete/update
Дата
Msg-id
p05111b10ba0872da9242@[192.168.1.4]
Ответ на
Re: performance of insert/delete/update (scott.marlowe)
Список
Дерево обсуждения
Re: [HACKERS] Realtime VACUUM, was: performance of insert/delete/update Jim Beckstrom <jrbeckstrom@sbcglobal.net>
Re: [HACKERS] Realtime VACUUM, was: performance of "scott.marlowe" <scott.marlowe@ihs.com>
>With Postgresql's implementation of MVCC, all your data are inserted in >real time, with a transaction date that makes the other clients ignore >them (mostly, other read committed transactions may or may not see them.) > >If there are indexes to update, they are updated in the same "invisible >until committed" way. > >All this means that your inserts don't block anyone else's reads as well. > >This means that when you commit, all postgresql does is make them visible. scott, Exactly the kind of explanation/understanding I was hoping for! Thank you! Tim
В списке pgsql-performance по дате отправления