[postgis-devel] GBOXF

[Mark Cave-Ayland]

On 30/11/11 11:51, Sandro Santilli wrote:

> What I mean is that whenever you use GROUP BY you must know you
> are grouping by bounding box equality and for that, it is well
> known that two very different geometries can have the very
> same bounding box. Seems unrelated to double precision vs. single
> precision to me.

Oh I appreciate that this exposes a limitation of GROUP BY for some use 
cases, but this behaviour is documented and predictable. By accepting 
double precision coordinates for the geometries, we are creating with 
the user that they will expect calculation to be done at similar 
precision. Therefore by rounding these values without the user's 
knowledge, we are breaking this contract and violating the principle of 
least surprise.

Actually while you mention it, I can also see how GROUP BY could be an 
issue with points in the same way as two points very close together 
could potentially be considered as one.


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