[postgis-tickets] [PostGIS] #4749: Improve behaviour of spatial predicates with GeometryCollection inputs

PostGIS trac at osgeo.org
Wed Sep 9 15:46:20 PDT 2020


#4749: Improve behaviour of spatial predicates with GeometryCollection inputs
--------------------------+---------------------------
  Reporter:  mdavis       |      Owner:  pramsey
      Type:  enhancement  |     Status:  new
  Priority:  medium       |  Milestone:  PostGIS 3.1.0
 Component:  postgis      |    Version:  2.5.x
Resolution:               |   Keywords:
--------------------------+---------------------------

Comment (by mdavis):

 Replying to [comment:4 pramsey]:
 > why even things like invalid multipolygons cannot be amenable to
 predicate evaluation?

 Can we say "GCs containing overlapping polygons"? "invalid" includes far
 too much craziness...

 > * A Intersects B... if any element A intersects any element of B, A
 intersects B

 Agreed

 > * A Contains B... if all elements of B are contained by at least one
 element of A, and are disjoint from all other elenents of A, then A
 contains B

 Some might say that "contains" should test against the entire area covered
 by the polygons (this is consistent with the notion of point-set
 topology).  In which case this can't be evaluated component-wise - it
 needs to use the (effective) union of the polygons.  But this semantic is
 certainly a subset of the full test.

 > * A Touches B... if all elements of A touch an element of B, and are
 disjoint from all other elements of B, then A touches B

 Sounds reasonable.

 > I mean, Relate() is not possible, but it feels like the boolean
 predicates, which are made up of some "obvious" interactions and other
 "ok, I can kind of believe that" intersection (do we really believe
 "crosses" is well defined between two polygons?) do have plausible rules
 that can be at least made up for GCs.

 So agreed, can probably make up some semantics which allow extension to
 all GC inputs.  I'm not sure the phone has been ringing off the hook to do
 this though?

 This ticket was really just about cleaning up how the current semantics
 are implemented.

-- 
Ticket URL: <https://trac.osgeo.org/postgis/ticket/4749#comment:5>
PostGIS <http://trac.osgeo.org/postgis/>
The PostGIS Trac is used for bug, enhancement & task tracking, a user and developer wiki, and a view into the subversion code repository of PostGIS project.


More information about the postgis-tickets mailing list