Re: Speeding up aggregates

Hannu Krosing kirjutas L, 07.12.2002 kell 02:32:
> Tom Lane kirjutas L, 07.12.2002 kell 01:46:
> > "Josh Berkus" <josh(at)agliodbs(dot)com> writes:
> > > What have other Postgres users done to speed up aggregates on large
> > > tables?
> >
> > FWIW, I've implemented hashed aggregation in CVS tip.
>
> Great!
>
> This should also make it easier to implement all kinds of GROUP BY
> ROLLUP|CUBE|GROUPING SETS|() queries.

Of these only ROLLUP can be done in one scan after sort, all others
would generally require several scans without hashing.

I just noticed that we don't even have a TODO for this. I think this
would be a good TODO item.

Bruce, could you add:

* Add ROLLUP, CUBE, GROUPING SETS options to GROUP BY

They are all defined in SQL99 p.79 <group by clause>

Some more background info (from a quick Google search)

a very short overview:
http://www.neddo.com/dm3e/sql3&olap.html

more thorough guide for DB2:
http://www.student.math.uwaterloo.ca/~cs448/db2_doc/html/db2s0/frame3.htm#db2s0279

-----------------
Hannu