Josh Berkus wrote:
> Tom,
> > I'm also kind of wondering if the main use-cases couldn't be met with
> > suitable multi-input custom aggregates, which is something we already
> > have as of 8.2.
> Actually, given that "skyline of" is *only* for aggregate sorting (as far as 
> I 
> can tell) it doesn't present the complications which QBE did for using a 
> function interface.   

There is people on a Venezuelan university working on SKYLINE OF and
other operators on Postgres.  I had some looks at their work because
they asked for help in the spanish list.

Not only they added the SKYLINE OF clause, but they also had some mods
to the ORDER BY clause, and a couple of other grammar changes as well.
While SKYLINE OF itself could probably be folded as aggregates, the
other stuff is not likely to be amenable to such treatment.

Also, keep in mind that there were plenty of changes in the executor.
This stuff is not likely to be very easy to implement efficiently using
our extant executor machinery; note that Ranbeer mentioned
implementation of "block nested loop" and other algorithms.  Not sure
how easy would be to fold that stuff into the optimizer for multi-input
aggregates, instead of hardwiring it to the SKYLINE OF syntax.

There's a certain group in the Venezuelan Uni that was about to finish
their thesis.  They promised me a look into their report; maybe I can
give further input from that and maybe merge Ranbeer's stuff with it.

Alvaro Herrera                      
PostgreSQL Replication, Consulting, Custom Development, 24x7 support

---------------------------(end of broadcast)---------------------------
TIP 3: Have you checked our extensive FAQ?


Reply via email to