Hello Arvind,
the cyclic change idea is really marvellous , thank you
You're welcome. :)
but we store each round, because we need player actions for further
analysis about game trends
Normally the different analyze-forms and goals are known from the
beginning. You could use a more compact format for the analysis, which
is optimized for this.
Also: if you just need this data for further analysis: don't store them
together with the actual game data! Big Performance NO GO!
Set up a second database-server! It can be a simple and even slow
machine. Cache the round-data compressed in RAM or another fast location
and import them from there in a steady flow into the second
database-server. Their you can do your heavy analyze at any time without
annoying your users!
(You can even use more than a second server to do the analyze in
parallel and sum up the aggregated data. Most the analyze must not be
actual, so it doesn't bother if it takes some hours more. It is a big
win for your user-performance)
Greetings from Germany,
Torsten
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
