Re: setFetchSize
От | Oliver Jowett |
---|---|
Тема | Re: setFetchSize |
Дата | |
Msg-id | CA+0W9LO5Q_icnHoYF8PyE6Aez0k__JjzAiREF=8=JWBZ8CxYsg@mail.gmail.com обсуждение исходный текст |
Ответ на | Re: setFetchSize (Tom Lane <tgl@sss.pgh.pa.us>) |
Список | pgsql-jdbc |
On 27 October 2011 12:18, Tom Lane <tgl@sss.pgh.pa.us> wrote: > Bottom line from a server-side point of view is that if you want that > overhead, you can ask for it, by opening a held cursor. It might be > sensible for JDBC to provide that functionality with something that > looks as much as possible like an ordinary resultset --- but I'm pretty > certain it shouldn't be the default behavior on the JDBC side either. We'd only do this if HOLD_CURSORS_OVER_COMMIT (which isn't the default) was explicitly set on the resultset. It really would be useful to specify something equivalent to WITH HOLD on a portal though - trying to work out how to transform an arbitrary query into a DECLARE CURSOR is the sort of SQL mangling the driver tries to avoid doing. (It did exactly that back in the 7.3 / protocol version 2 days and there were all sorts of nasty edge cases) Oliver
В списке pgsql-jdbc по дате отправления: