> ------------ Původní zpráva ------------
> Od: Andreas Kretschmer <akretschmer@spamfence.net>
> Předmět: Re: [GENERAL] postgreSQL multithreading
> Datum: 30.3.2008 17:39:02
> ----------------------------------------
> postgre@seznam.cz <postgre@seznam.cz> schrieb:
>
> > Hi all,
> >
> > I was trying to find some way to implement multithreading into my postgreSQL
> stored functions.
> >
> > The thing is, that I have data stored in multiple tables - for each day one
> table - and I want to write
> > a function which selects data from these tables and stores them into files (or
> just returns the data in cursor);
> > Information about data stored in these tables are in another table defined
> like this:
> >
> > partitions_daily(from_date timestamp, to_date timestamp, table_name)
> >
> > From_date and to_date tells in which table the requested data should be
> found.
> >
> > In this time I have function through which I can get the data, the problem is
> that everything is processed with only
> > one CPU core. I think it should be very easy to make this model to be
> multithreaded and use more CPU cores.
>
> I think, the only way to achieve this is to use multiple connections to
> the database. One connection to select the table_name and than, for
> every table_name, a own connection to the database to select and write
> the data to files.
>
>
What about some C dynamically loaded function in which I could call new connection for each thread?
Lukas Houf