Re: Intersection tests with curved polygons

These are literally CurvePolygon type? It’s probably getting caught in our lack of full curve support.
I would be interested in the ST_Distance between the point and those two CurvePolygons. (Because, for distance, we have a postgis-native implementation that supports curves). Whereas for intersection, the calculation is delegated to GEOS *after linearizing the inputs*. In that linearization, could sit the logically problem you’re seeing.
P

> On Dec 20, 2024, at 5:05 AM, Andrea Aime <[email protected]> wrote:
> 
> Hi,
> I'm having an issue testing the intersection between a point and two curved polygons.
> Here is the situation visually, from QGIS:
> 
> <image.png>
> 
> and zooming in (a lot):
> 
> <image.png>
> 
> In particular, the point is (barely) inside the polygon at the top, whose identifier is 12875.
> Now if I run an intersection test, I get both polygons as a result:
> 
> > SELECT ogc_fid FROM testdata WHERE ST_Intersects(geom, ST_GeomFromText('POINT (25492818 6677399.98)', 3879));
>  ogc_fid
> ---------
>     1258
>    12875
> (2 rows)
> 
> As you can see... both polygons tested positive for intersection. Visually that does not make sense, the point is in a hole inside polygon 1258, but maybe there is tolerance/linearization in play, or it's just a visualization fluke (how the curved polygons are rendered)... but I've also visualized the same data in GeoServer, and the visual result is the same.
> 
> I have verified the behavior in the following two version:
> postgis 3.2 and postgresql 14.15
> postgis 3.5 and postgresql 17.2 (latest postgis docker image)
> Test data can be found in this Github GIST <https://gist.github.com/aaime/dcf55b635c5bd09c2b9ebb31a2ab083a>.
> 
> Can anyone shed some light on the behavior I'm seeing?
> 
> Best regards
> Andrea