[postgis-users] spatial index for topology?

[William Kyngesburye]

On Oct 27, 2013, at 5:22 AM, Sandro Santilli wrote:

> On Sat, Oct 26, 2013 at 06:27:41PM -0500, William Kyngesburye wrote:
>> Is there any way to spatially index topogeometry?
> 
> Nope, due to the fact that TopoGeometry spatial extent is not
> in the value itself but derived using a volatile function from
> external (referenced) data.
> 
> This is a known problem:
> http://trac.osgeo.org/postgis/ticket/2124
> http://trac.osgeo.org/postgis/ticket/853
> 
> One idea would be to make the && operator be
> implemented by an SQL-defined function to be
> fully inlined and thus giving the planner enough
> hints to use the index on the underlying primitive
> tables, but feasibility of that is uncertain.
> 
> Another idea would be to introduce a trigger-based
> mechanism to update some form of bbox cache into
> each TopoGeometry affected by any change into any
> primitive table, and then use that cache in select
> querys. Note that making it transparent (using &&
> operator) would require a change in the TopoGeometry
> structure, to add a bounding box field. 
> 
> --strk;

OK.

So for now, working with a large topology data set is not practical.  Some processing should not be a problem, if it operates on the primitives, which are indexed.

-----
William Kyngesburye <kyngchaos*at*kyngchaos*dot*com>
http://www.kyngchaos.com/

Earth: "Mostly harmless"

- revised entry in the HitchHiker's Guide to the Galaxy