[postgis-devel] Death to Pointless Operators

[Mark Cave-Ayland]

On 25/12/10 16:38, Paragon Corporation wrote:

>> The default gserialized index is now 2d and uses the full set of old
> operators.
>
> Paul,
> Does this mean we have to choose at compile time whether we get a 3D or 2D
> index.  That would be a big bummer.

No. What I was suggesting was to create a separate operator class for 
the 3D index for people that want to use it. For example, at the moment 
you create a 2D index like this:

CREATE INDEX [indexname] ON [tablename] USING GIST ( [geometryfield] );

I was suggesting that we create a non-default operator class so that 
people who want ND indices can create them using something like this:

CREATE INDEX [indexname] ON [tablename] USING GIST ( [geometryfield] 
gist_geometry_nd_ops );

Hence we can still offer a migration path for existing PostgreSQL R-Tree 
users and the existing index semantics can be used if people find that 
the new indices break their spatial world.

> Is the idea of having an&&&  (a 3d interacts operator) out of the question
> if we have 2D operators?

It depends on the exact behaviour of the operators and operator classes. 
I think probably the best thing to do here would be to raise a new 
thread on -hackers, let them know what we are trying to do, and then let 
them point us in the right direction.


ATB,

Mark.

-- 
Mark Cave-Ayland - Senior Technical Architect
PostgreSQL - PostGIS
Sirius Corporation plc - control through freedom
http://www.siriusit.co.uk
t: +44 870 608 0063

Sirius Labs: http://www.siriusit.co.uk/labs