On Monday 05 November 2001 03:24, Tille, Andreas wrote:
> I discussed the issue of using statistics tables to speed up certain
> queries. He told me that those technique is known as OLAP tubes in
> MS SQL server and that there are tools to build such things. Is this
> a valid comparison? He did not use it because it would disable the
> access solution of our clients. Are there any tools for PostgreSQL for
> such stuff besides the manual creating tables and triggers?
I still don't understand your guy. Knowing that the table (and with it the
performance demands) will grow, it is quite stubborn and certainly not
elegant at all to insist on the blunt query instead of a smart solution. The
smart solution as outlined always returns results instantly and needs next to
no memory or other ressources as compared to the blunt query, regardless of
the growth of your database. It would only impact the growth *rate* due to
the fired triggers, but then, your application does not seem to have a heavy
insert load anyway and you could always queue the inserts with middleware as
you have no realtime demands.
Btw, what is wrong with creating a few tables and a few trigger functions
"manually"? Writing, testing, and debugging them should not cost more than
a couple of days. Why would I want a tool for it? I might spend a couple of
hours writing a python script if I would need similar triggers for many
tables over and over again, but your problem does not seem to have the need
for this.
Horst