Re: [HACKERS] curso$B#r(Bs in LLL

Поиск
Список
Период
Сортировка
От Vadim Mikheev
Тема Re: [HACKERS] curso$B#r(Bs in LLL
Дата
Msg-id 36197B56.3949B470@krs.ru
обсуждение исходный текст
Ответ на cursors in LLL  ("Hiroshi Inoue" <Inoue@tpf.co.jp>)
Ответы RE: [HACKERS] cursors in LLL
Список pgsql-hackers
Hiroshi Inoue wrote:
>
> Hi all.
> I'm looking forward to the appearance of LLL in PostgreSQL 6.5 and have a
> question about the sensitivity of cursors in LLL.
>
> In LLL cursors are INSENSITIVE as Oracle ?
>
> Currently cursors are indeterminate and in some cases they are strangely
> sensitive(for me).

Do you mean seeing row inserted between fetches ?
Should this be changed ?
How is this in Oracle, Informix, Sybase, standards ?

> In LLL the behavior of cursors will be more complicated, if changes by other
> transactions can be seen by fetch statements(especially for read committed
> isolation level).
>
> I hope INSENSITIVE cursors to be implemented whose behavior we can predict
> and I think that they can be realized according to proposals for LLL by
> Vadim.
>
> In LLL access methods return snapshot of data as they were in _some_ point
> in time.
> For read committed mode this moment is the time when statement began.
> For serializable mode this is the time when current transaction began.
>
> For a INSENSITIVE cursor this is the time when it was opened(declared),
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This is easy to implement.
But I'd like to know what standards say about cursor sensitivness...

> not the time when the fetch statements for it began ?

Thanks.

Vadim

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Tatsuo Ishii
Дата:
Сообщение: Re: [HACKERS] Open 6.4 items
Следующее
От: Bruce Momjian
Дата:
Сообщение: Re: [HACKERS] RE: [GENERAL] Long update query ? (also Re: [GENERAL] CNF vs. DNF)