[postgis-users] Problems with st_distance
Stuart Guthrie
stuartguthrie at gmail.com
Tue Mar 13 14:37:33 PDT 2012
Thanks Nicklas and Charles for your help. I'm a babe in the woods with this
even after reading the postgis book. Lots of new terminology.
I'll check your information today and report back to the list.
ST_Distance(geog,geog) is supposed to report only in meters according to
the documentation.
http://www.postgis.org/documentation/manual-svn/ST_Distance.html
"For geography type defaults to return the minimum distance around WGS 84
spheroid between two geographies in meters. "
I'm thinking it' a planar vs spheriod thing. I'll do the same as Nicklas
and check on Google Earth.
The trick with GIS is that it's not like picking up a new API set, you've
got a whole world of scientific terms to start understanding!
Again, thanks very much for your help. Great list!
On Tue, Mar 13, 2012 at 11:26 PM, Nicklas Avén
<nicklas.aven at jordogskog.no>wrote:
> Hallo again
>
> Stuart, what makes you say that you know the distance is 1800 meters?
> I just measured by hand in google earth and I get the same answer as
> PostGIS there about 1180 meters.
>
> I think you have been tricked by some application using web mercator
> projection in the first.
>
> Google earth, like PostGIS geography knows how to calculate distances on
> a sphere or spheroid which for instance google maps doesn't do since it
> is a flattened globe.
>
> HTH
>
> Nicklas
>
>
>
>
>
>
> On Tue, 2012-03-13 at 12:50 +0100, Nicklas Avén wrote:
> > Sorry, wrong from me
> >
> > Something is wrong of course if the answer is 1179 compared to 1800. My
> > brain was on vaccation
> >
> > /Nicklas
> >
> >
> > On Tue, 2012-03-13 at 12:48 +0100, Nicklas Avén wrote:
> > > The geography type distance calculations gives answers in meters from
> > > unprojected data (4326).
> > >
> > > I see no erason to belive that the postgis answer is wrong, especially
> > > not when Stuart says he knows the distance is 1800 meters and he gets
> > > 1179.61 meters from PostGIS.
> > >
> > > I don't know anything about that draft-logic site but I would definitly
> > > rely more on the PostGIS answer.
> > >
> > > The measurments on Bing, Google maps and so on will be wrong since the
> > > webmercator projection 913900 is a flattened globe. How could it be
> > > correct?
> > >
> > > /Nicklas
> > >
> > >
> > > On Tue, 2012-03-13 at 07:37 -0400, Charles Galpin wrote:
> > > > The distance should be in the projection units, which for 4326 is
> > > > decimal degrees, and if you want meters you'll need to convert the
> > > > results to meters.
> > > >
> > > >
> > > > But first, I suspect your data is not in 4326 then if you get 1179.
> > > > Make sure you are using the SRID when inserting them
> > > >
> > > >
> > > > > update address set geog=ST_GeographyFromText('POINT('||
> > > > > lon||' '||lat||')',4326) where id=?;
> > > > Also see
> > > > http://postgis.refractions.net/docs/ST_Distance.html
> > > > hth
> > > > charles
> > > >
> > > > On Mar 13, 2012, at 2:58 AM, Stuart Guthrie wrote:
> > > >
> > > > > Could it be returning a distance in decimal miles? I'm out by
> around
> > > > > 1.6 - 1.6km = 1mile?
> > > > >
> > > > > On Tue, Mar 13, 2012 at 5:43 PM, Stuart Guthrie
> > > > > <stuartguthrie at gmail.com> wrote:
> > > > > OK, I'm a newbie to postgis but an old-timer with open
> > > > > source in general.
> > > > >
> > > > > Sorry to bother everyone but this problem is driving me a
> > > > > little nuts.
> > > > >
> > > > > I've bought the 'postgis in action' book and now I think I
> > > > > still know nothing.
> > > > >
> > > > > I'm trying to store long lats and do queries on them to
> > > > > establish distance and use that for reporting.
> > > > >
> > > > > I'm using this to update the table 'address'
> > > > >
> > > > > update address set geog=ST_GeographyFromText('POINT('||
> > > > > lon||' '||lat||')') where id=?;
> > > > >
> > > > > the address table field 'geog' has this definition:
> > > > >
> > > > > geog geography(Point,4326)
> > > > >
> > > > > When I compare two addresses that I know are around 1800m
> > > > > apart I get this:
> > > > >
> > > > > select a.addressline2, a.suburb, ST_Y(a.geog::geometry),
> > > > > ST_X(a.geog::geometry),
> > > > > b.addressline2,b.suburb,ST_Y(b.geog::geometry),
> > > > > ST_X(b.geog::geometry), ST_distance(a.geog,b.geog ) from
> > > > > address a join address b on b.id=41491 where a.id=1238 ;
> > > > >
> > > > > addressline2 | suburb | st_y | st_x |
> > > > > addressline2 | suburb | st_y | st_x |
> > > > > st_distance
> > > > >
> ---------------+--------+-------------+-------------+----------------+-------------+-------------+-------------+------------------
> > > > > 140 Sussex St | Sydney | -33.8692429 | 151.2036312 | 1
> > > > > Riley Street | Surry Hills | -33.8726426 | 151.2157115 |
> > > > > 1179.61881540556
> > > > >
> > > > > Postgis says they are 1179.61m apart.
> > > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > postgis-users mailing list
> > > > postgis-users at postgis.refractions.net
> > > > http://postgis.refractions.net/mailman/listinfo/postgis-users
> > >
> > >
> > > _______________________________________________
> > > postgis-users mailing list
> > > postgis-users at postgis.refractions.net
> > > http://postgis.refractions.net/mailman/listinfo/postgis-users
> > >
> >
> >
> > _______________________________________________
> > postgis-users mailing list
> > postgis-users at postgis.refractions.net
> > http://postgis.refractions.net/mailman/listinfo/postgis-users
> >
>
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>
--
Best regards
Stuart Guthrie
(m) 0403 470 123
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20120314/3dc26c3c/attachment.html>
More information about the postgis-users
mailing list