[postgis-users] The operator && is not overlaps operator based onthe OGC SF SQL

[Nicolas RIBOT]

Hello Darko,

No. && overlap operator is not related to OGC definition of spatial
operators:
The "&&" operator works on geometries' bounding box (extent), not on the
geometries themselves.

"&&" tells if 2 geometries' extents overlap or not. Spatial index uses
it to exclude geometries from a spatial query, speeding up the process.

OGC spatial operators are implemented in the GEOS extension and follow
the semantic you explained.
They also follow the OGC naming convention: overlap(geometry, geometry),
intersect(geometry, geometry), etc...

Nicolas

> 
> Hello,
> 
> I am new user of the PostGIS and I have just read the PostGIS 
> manual. I
> found that in order to use the spatial index one should use the "&&"
> operator.
> This operator ("&&") is named the "overlap" operator (page 6, 
> page 14 of the
> manual).
> I am not sure whether this definition is related to the 
> "overlaps" operator
> in the OGC Simple Features Specification For SQL or not.
> In the OGC doc, the "overlaps" operator is defined only for 
> A/A L/L and P/P
> and it is defined like:
> 
> a.Overlaps(b)<=>(dim(I(a)) = dim(I(b)) = dim(I(a) ^ I(b))  ^  
> (a^b != a) ^
> (a^b != b)
> 
> If that is the case, then the postgis "&&" operator means actually
> "intersects" and not "overlaps".
> I tried this by executing a few query and found that the 
> operator "&&" might
> be an implementation of the "intersect" operator.
> 
> I would appreciate some comments and clarifications.
> 
> Regards,
> Darko
> 
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>