Re: ECPG support for PQsetChunkedRowsMode()
| От | Tim Fors |
|---|---|
| Тема | Re: ECPG support for PQsetChunkedRowsMode() |
| Дата | |
| Msg-id | CAOjiQ0C8O4g_KcVmOMq_+WJy+YCugBYe8A=RsZgHGP1nfhQG=g@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: ECPG support for PQsetChunkedRowsMode() ("Daniel Verite" <daniel@manitou-mail.org>) |
| Ответы |
Re: ECPG support for PQsetChunkedRowsMode()
|
| Список | pgsql-interfaces |
Thanks for the info Daniel.
The reason I'm asking in the context of ECPG is this: When using a cursor to FETCH one row of a result set at a time, it seems that the server literally sends just one record to the client at a time, so using a cursor in this context results in a lot of communication between the client and server in order to process a large number of rows. I'm wondering if this is accurate and if so, is there a way to improve performance by enabling some kind of buffering or caching on the client side, while still allowing the cursor to function as normal.
Note that ECPG supports using a C array as a host variable, and in that case there's a bulk transfer of rows from the client to the server, so performance is much better than getting the rows one at a time using a cursor. However, AFAIK this is not standard coding practice - one expects to be able to use the cursor to FETCH one row at a time, but still offer good performance.
When I saw the addition of this "chunked rows mode" I thought perhaps it would enable faster cursor usage. Since your answer indicates that it would not, do you have any other recommendations to limit the effect of network latency in order to speed up retrieval of rows via a cursor?
Thanks,
Tim
On Wed, Nov 19, 2025 at 1:44 PM Daniel Verite <daniel@manitou-mail.org> wrote:
Tim Fors wrote:
> The "chunked rows mode" appears to have been added to
> libpq in V17.
>
> Does ECPG support this mode
No. It's not clear how it would fit. This mode implies the use
of PQsendQuery(), which does not seem to have an equivalent in
ECPG.
> are there plans to do so in the near future?
Postgres as a project does not have a roadmap, but
what's being worked on at any point in time can
generally be tracked at https://commitfest.postgresql.org
> Also, is there any data available on performance improvements that
> one might see by using this mode?
In [1], the time to retrieve rows by psql was compared between the
the chunked rows mode and the cursor-based implementation, the
latter being what would typically be used by an ECPG app.
[1]:
https://postgr.es/m/202bae70-f60d-4d24-a8d8-50b4ed638887@manitou-mail.org
Best regards,
--
Daniel Vérité
https://postgresql.verite.pro/
В списке pgsql-interfaces по дате отправления: