[postgis-devel] [PostGIS] #1976: ST_Intersects() returns false while ST_Intersection returns a geography

[PostGIS]

#1976: ST_Intersects() returns false while ST_Intersection returns a geography
------------------------+---------------------------------------------------
  Reporter:  mhiper3pg  |       Owner:  pramsey      
      Type:  defect     |      Status:  closed       
  Priority:  medium     |   Milestone:  PostGIS 1.5.6
 Component:  postgis    |     Version:  1.5.X        
Resolution:  invalid    |    Keywords:               
------------------------+---------------------------------------------------

Comment(by mhiper3pg):

 I see how ST_Intersects() and ST_Intesection() may yield different
 restuls. However, I am not so sure that ST_Intersects() is returning the
 correct result here (sorry if I am being obtuse).

 I see the issue with the flat Earth, as I also have several of those
 cases. However, what I thought weird in this case is that I had expected
 the great circles to actually "expand" the shape of the north border,
 which is the closest one, and the one that is defined only by a small set
 of points (and therefore, more likely to have long straight lines in the
 linear interpolation). In that case, the small polygon should still
 intersect with the state.

 Another issue is that, when I find this situation, I can usually extend
 the small polygon in the direction of the closest border of the large
 polygon, and ST_Intersects() will still return false (meaning that the
 small polygon is "outside" the large one). However, I cannot do so in this
 case. (I realize this "test" may fail depending on the shape of the large
 polygon, but I don't see how it would in this case).

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/1976#comment:2>
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.