Re: 2D partitioning of VLDB - sane or not?

Jason,

> Aside from running into a known bug with "too many triggers" when creating
> gratuitous indices on these tables, I feel as it may be possible to do what
> I want without breaking everything. But then again, am I taking too many
> liberties with technology that maybe didn't have use cases like this one in
> mind?

Well, you're pushing PostgreSQL partitioning further than it's currently able
to go. Right now our range exclusion becomes too costly to be useful
somewhere around 300 to 1000 partitions (depending on CPU and other issues)
because the constraints are checked linearly.

To make your scheme work, you'd need to improve our existing partitioning
mechanisms to scale to 100,000 partitions. It would also help you to
implement multiple inheritance so that you could have a partition which
belonged to two masters. I'd be very interested in seeing you do so, of
course, but this may be more hacking than you had in mind.

--
Josh Berkus
PostgreSQL @ Sun
San Francisco