[postgis-devel] [PostGIS] #230: Put in costs in _ST and possibly other functions
PostGIS
trac at osgeo.org
Mon Aug 10 02:56:44 PDT 2009
#230: Put in costs in _ST and possibly other functions
----------------------+-----------------------------------------------------
Reporter: nicklas | Owner: robe
Type: defect | Status: assigned
Priority: medium | Milestone: postgis 1.5.0
Component: postgis | Version: 1.4
Resolution: | Keywords: st_dwithin st_expand bbox
----------------------+-----------------------------------------------------
Comment (by mcayland):
Yes, correct. The aim should be for us to pick defaults that "just work"
for the majority of people rather than complicating matters by having to
tune individual installations.
I know from my work on the statistics side of PostGIS that it is a
pointless exercise trying to perfect our estimates so that they are
accurate to reality; this is mainly because the existing routines in
PostgreSQL are only estimates too. All we need to do is ensure that the
more expensive functions are weighted heavier than their cheaper
equivalents by a reasonable amount of units so given a dead heat in terms
of planner cost, it will always choose the cheaper function first.
ATB,
Mark.
--
Ticket URL: <http://trac.osgeo.org/postgis/ticket/230#comment:22>
PostGIS <http://trac.osgeo.org/postgis/>
PostGIS
More information about the postgis-devel
mailing list