Re: cursors outside transactions
| От | Alvaro Herrera |
|---|---|
| Тема | Re: cursors outside transactions |
| Дата | |
| Msg-id | 20030318030156.GB6116@dcc.uchile.cl обсуждение исходный текст |
| Ответ на | cursors outside transactions (Neil Conway <neilc@samurai.com>) |
| Ответы |
Re: cursors outside transactions
|
| Список | pgsql-hackers |
On Mon, Mar 17, 2003 at 09:48:34PM -0500, Neil Conway wrote: > (2) Use MVCC to ensure that the snapshot of the database that the > transaction had is still valid, even after the transaction itself has > committed. What about opening a pseudo-transaction that exists only to serve the cursor? That frees you from modifying VACUUM and resource management. The transaction should be commited (aborted?) when the cursor is closed. Maybe you can call the lowlevel transaction routines directly. -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "Postgres is bloatware by design: it was built to house PhD theses." (Joe Hellerstein, SIGMOD annual conference 2002)
В списке pgsql-hackers по дате отправления: