Re: Temporary Views
От | Bruce Momjian |
---|---|
Тема | Re: Temporary Views |
Дата | |
Msg-id | 200208131531.g7DFVNT07945@candle.pha.pa.us обсуждение исходный текст |
Ответ на | Re: Temporary Views (Rod Taylor <rbt@zort.ca>) |
Ответы |
Re: Temporary Views
|
Список | pgsql-hackers |
Rod Taylor wrote: > On Tue, 2002-08-13 at 11:18, Bruce Momjian wrote: > > Tom Lane wrote: > > > Gavin Sherry <swm@linuxworld.com.au> writes: > > > > I'm thinking that temporary views should be pretty trivial to > > > > implement. > > > > > > ... except not so trivial, per the rest of your note. > > > > > > Do we actually need any such feature? Views on temp tables already work > > > correctly in CVS tip: the implicit DROP CASCADE on temp tables at > > > backend exit makes such views go 'way too. > > > > Oh. but RESTRICT is the default. Seems like the view should go away no > > matter what, and if they mix temp and non-temp tables, is it obvious > > that the view will disappear if they didn't specify TEMP on view > > creation. > > When the backend exits the code that removes temp tables is CASCADE by > default and anything depending on it will disappear. Oh, OK, that is interesting. So that only leaves the issue of not specifying TEMP in a case of views using mixed temp/non-temp tables. We don't specify TEMP when creating an index on a temp table, and it is auto-destroyed. I guess it is OK that we don't specify TEMP on a view creation using a temp table, except that the view can have a mix of temp and non-temp while an index is just on one table. I can go either way on this. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania19073
В списке pgsql-hackers по дате отправления: