[postgis-users] ST_Overlaps broken ?

[Mike Toews]

On 11 September 2013 10:56, Jean Marchal <jean.d.marchal at gmail.com> wrote:
> Dear PostGIS users,
>
> <snip>
>
> Returns FALSE instead of TRUE, same thing for hundreds of polygons. Is it
> just me or everyone get same result ?
>
> I tried with postgis 2.1 and 2.1.1dev
>
> My guess is it's not only ST_Overlaps() but some underlying function called
> by ST_Overlaps()

This is the expected result. It returns true for ST_Covers, which
means one of the geometries is completely contained in the other, so
by definition it is false for ST_Overlaps. Maybe you are looking for a
different spatial predicate? I'd recommend putting these geometries in
JTS test builder to get a visual reference for how the spatial
predicates actually work. Also see
https://en.wikipedia.org/wiki/DE-9IM

-Mike