[postgis-tickets] [PostGIS] #2422: geography regression difference ST_DWithin

[PostGIS]

#2422: geography regression difference ST_DWithin
---------------------+------------------------------------------------------
 Reporter:  robe     |       Owner:  pramsey      
     Type:  defect   |      Status:  new          
 Priority:  high     |   Milestone:  PostGIS 2.1.1
Component:  postgis  |     Version:  2.1.x        
 Keywords:           |  
---------------------+------------------------------------------------------

Comment(by robe):

 actually never mind the geometry collection bit (its a mix of polygons and
 geometry collections).  If my 2 geometries that failed this

 {{{
 SELECT gid, geog FROM  face
 WHERE NOT _ST_DWithin(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography, 1609.344,true )
 AND _ST_DWithinUnCached(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography, 1609.344 )

 }}}

  and create a smaller only 2 row table with it.  Then only one fails this
 time but putting the check in the from yield2 true so definitely seems
 like a caching bug of sorts.


 {{{
 SELECT gid, geog, _ST_DWithin(geog, st_setsrid(st_makepoint(-71.0636,
 42.3584), 4326)::geography, 1609.344,true ) FROM  face2
 WHERE NOT _ST_DWithin(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography, 1609.344,true )
 AND _ST_DWithinUnCached(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography, 1609.344 )
 }}}

 Yields:

 {{{
 gid | _st_dwithin
 ----+-------------
   2 | t
 }}}


 {{{
 SELECT gid, _ST_DWithin(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography, 1609.344,true )
 , ST_Distance(geog, st_setsrid(st_makepoint(-71.0636, 42.3584),
 4326)::geography )
 FROM  face2;

  gid | _st_dwithin |  st_distance
 -----+-------------+----------------
    1 | t           | 1398.407057514
    2 | f           |   1001.9245264
 }}}

 Berry berry interesting.  What's kinda bizarre to me is both records ARE
 within distance so even caching wise, how could it ever have a cache that
 has false in it.

 I still need to upgrade this db to test again.  This is running with
 released 2.1.0

-- 
Ticket URL: <http://trac.osgeo.org/postgis/ticket/2422#comment:23>
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.