[postgis-devel] [PostGIS] #1625: ST_Within and ST_DWithin disagree

[PostGIS]

#1625: ST_Within and ST_DWithin disagree
---------------------+------------------------------------------------------
 Reporter:  strk     |       Owner:  pramsey      
     Type:  defect   |      Status:  new          
 Priority:  medium   |   Milestone:  PostGIS 2.0.0
Component:  postgis  |     Version:  trunk        
 Keywords:           |  
---------------------+------------------------------------------------------

Comment(by nicklas):

 One way to quantify how much the two algorithms disagree is to see how
 much the point have to be buffered to intersect with the line. The answer
 seems in this case to be 0.0000000001 units which is far less than the
 precision of the input.


 {{{
 select
 st_intersects('010200000002000000B6813B20D7F92041F5807988B8AF014100000000D6F9204133333333BBAF0141'::geometry,st_buffer('010100000070841C1ED7F9204187A97F8DB8AF0141'::geometry,
 0.0000000001));
 }}}

 This returns true


 I have twisted my head when I was looking at #1502 to understand what is
 right and what is wrong. I don't know, but to get consistent answers
 beyond the precision of the input I guess is only possible if the approach
 is identical (so the error is identical).

 Does that make sense?

 My conclusion is that we might have to start thinking about how to handle
 tolerance and precision in generic terms.

 From what I think SQL-Server is outputting their answers with far less
 precision than double precision and in ArcGIS world you tell the system
 what precision you want to use.

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/1625#comment:6>
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.