On Thu, 2024-01-25 at 04:46 +0000, PG Doc comments form wrote:
> The following documentation comment has been logged on the website:
>
> Page: https://www.postgresql.org/docs/16/protocol-flow.html
>
> Please consider the following exchange with a PG database (Kotlin + Reactor+
> r2dbc-postgresql):
>
> [trace of a pipelined message flow]
>
> The application code is the following (for the sake of clarity):
>
> [some Java code]
>
> I’m aware that Postgres doesn’t support cursor(portal) multiplexing so its
> impossible to fetch data for two selects simultaneously but should I make
> use of selects pipelining when I want to make the most of the DB? I mean to
> say does a single Postgres backend employs concurrent processing and its
> possible for a DB to transmit results for a query #2 (see the traffic) and
> concurrently parse/rewrite/plan/optimize the subsequent (pipelined) queries
> (issued from the same transaction). Neither chapter 55 nor 52 expand on it
> (or it’s hard to spot at least).
No, PostgreSQL is single-threaded (with the exception of parallel query execution).
Yours,
Laurenz Albe