Re: DELETE with LIMIT (or my first hack)
От | Daniel Loureiro |
---|---|
Тема | Re: DELETE with LIMIT (or my first hack) |
Дата | |
Msg-id | AANLkTik9F9sm_o-F76Go1QdTkLK8Jn3d5QzNtE86ctZf@mail.gmail.com обсуждение исходный текст |
Ответ на | DELETE with LIMIT (or my first hack) (Daniel Loureiro <loureirorg@gmail.com>) |
Ответы |
Re: DELETE with LIMIT (or my first hack)
Re: DELETE with LIMIT (or my first hack) |
Список | pgsql-hackers |
> 3. This doesn't work tremendously well for inheritance trees, where<br /> > ModifyTable acts as sort of an implicitAppend node. You can't just<br /> > funnel all the tuples through one Sort or Limit node because they aren't<br/> > all the same rowtype. (Limit might perhaps not care, but Sort will.)<br /> > But you can't have a separateSort/Limit for each table either, because<br /> > that would give the wrong behavior. Another problem with funnelingall<br /> > the rows through one Sort/Limit is that ModifyTable did need to know<br /> > which table eachrow came from, so it can apply the modify to the right<br /> > table.<br /><br />So I guess that I have choose thewrong hack to start.<br /><br />Just for curiosity, why the result of "WHERE" filter (in SELECT/DELETE/UPDATE) is notput in memory, i.e. an array of ctid, like an buffer and then executed by SELECT/DELETE/UPDATE at once ?<br /><br />Greets,<br/>--<br />Daniel Loureiro<br />
В списке pgsql-hackers по дате отправления: