> > How were you inserting the data? Were you doing multiple inserts per
> > transactions? Copy? That sounds really slow to me.
>
> The database mainly contains integers, which represent the behavior of
> internet users on our site, so it's very compact data.
> We used multiple insert with mysql but i did'nt find that option in postgres.
BEGIN;
INSERT INTO table (val1, val2, val3) VALUES (1,2,3);
INSERT INTO table (val1, val2, val3) VALUES (4,5,6);
INSERT INTO table (val1, val2, val3) VALUES (7,8,9);
INSERT INTO table (val1, val2, val3) VALUES (10,11,12);
[...]
COMMIT;
> > > By testing we created a postgres database to on an other server
> > > (same type). The copy command did'nt work, because of 'strange
> > > characters', so we used normal inserts. It took us 12 hours to
> > > import, and 10 hours to create the indexes.
> >
> > Have you tried to escape the data before you inserted it? That
> > should've solve things.
>
> No, how do you do that ?
Do you know what those 'strange characters' are? -sc
--
Sean Chittenden
