[postgis-users] Coordinates and SRID - bound overflow, ST_Equals

Thu Jan 20 07:02:19 PST 2011

Doug,

There are cases where we need to use a difference branch of longitude values, e.g. 0 to 360, rather than -180 to 180.  For example, if one has a polygon that spans the international dateline, and one uses longitude from -180 to 180, some vertices of the polygon will have longitudes near 180, while others will have longitudes near -180.  When drawn as a Cartesian GEOMETRY, this will contain lines that span the entire earth.  

To work around the limitation of Cartesian geometries for polygons and lines, we sometimes use longitudes from 0 to 360 and other times use longitudes from -180 to 180, depending on where the geometries are located.  Because our geometries span the globe, we cannot use either of these branches of longitude exclusively.  

We want to convert these geometries to geographies to avoid the Cartesian limitation, and it would be nice if the transformation could be automatic through the casting function.

Radu

From: postgis-users-bounces at postgis.refractions.net [mailto:postgis-users-bounces at postgis.refractions.net] On Behalf Of Doug_Newcomb at fws.gov
Sent: Thursday, January 20, 2011 9:23 AM
To: PostGIS Users Discussion
Subject: Re: [postgis-users] Coordinates and SRID - bound overflow, ST_Equals

Radu, 
        I am somewhat confused.  From my somewhat limited understanding of EPSG:4326 it's limited to values between 180, -180 for longitude and 90,-90 for latitude.  If your source geometry is epsg:4326, there should not be values outside of those limits.  Are you wrapping around those limits in your geometry column to accomodate lines/polygons  that cross those boundaries ? ( Stretching my brain a bit here)  Can you give examples of coordinates that you have that are exceed these limits?   

Doug 

Doug Newcomb             
USFWS
Raleigh, NC
919-856-4520 ext. 14 doug_newcomb at fws.gov
---------------------------------------------------------------------------------------------------------
The opinions I express are my own and are not representative of the official policy of the U.S.Fish and Wildlife Service or Dept. of the Interior.   Life is too short for undocumented, proprietary data formats. 

"Ilie, Radu" <rilie at wsi.com> 
Sent by: postgis-users-bounces at postgis.refractions.net 

01/20/2011 08:33 AM 

Please respond to
PostGIS Users Discussion <postgis-users at postgis.refractions.net>

To

"PostGIS Users Discussion" <postgis-users at postgis.refractions.net> 

cc

Subject

Re: [postgis-users] Coordinates and SRID - bound overflow, ST_Equals

Doug, 

We do have the same problem. We are in a situation where a central server has a column of type geometry (SRID 4326) and it has points with latitude/longitude outside of (-180 -90,180 90). We need to retrieve the values from this column and insert them into a local server where the column is of type geography. It doesn’t work, we get errors like this: 

ERROR:  Coordinate values are out of range [-180 -90, 180 90] for GEOGRAPHY type 

Our solution for now is to modify the PostGIS code and normalize points outside of (-180 -90,180 190) during the cast to geography. 

The solution is cumbersome, because it requires us to build PostGIS from source under Windows (the local server is Windows 7). And that is a science project in itself. 

For reasons outside our control we cannot change the type of the column on the central server, nor can we move the local server to Linux. Is there any way to use the existing functions in PostGIS to normalize points outside the allowed range for geography? We need something that can be used during an implicit cast. 

Radu Ilie 

From: postgis-users-bounces at postgis.refractions.net [mailto:postgis-users-bounces at postgis.refractions.net] On Behalf Of Doug_Newcomb at fws.gov
Sent: Thursday, January 20, 2011 8:02 AM
To: PostGIS Users Discussion
Subject: Re: [postgis-users] Coordinates and SRID - bound overflow, ST_Equals 

Balint 
Would using the geography data type be more suitable to what you are trying to accomplish? 

http://postgis.refractions.net/docs/ch04.html#PostGIS_GeographyVSGeometry <http://postgis.refractions.net/docs/ch04.html#PostGIS_GeographyVSGeometry>  
Doug Newcomb             
USFWS
Raleigh, NC
919-856-4520 ext. 14 doug_newcomb at fws.gov
---------------------------------------------------------------------------------------------------------
The opinions I express are my own and are not representative of the official policy of the U.S.Fish and Wildlife Service or Dept. of the Interior.   Life is too short for undocumented, proprietary data formats. 

Balint Persics <persicsb at gmail.com> 
Sent by: postgis-users-bounces at postgis.refractions.net 

01/20/2011 06:30 AM 

Please respond to
PostGIS Users Discussion <postgis-users at postgis.refractions.net>

To

postgis-users at postgis.refractions.net 

cc

Subject

[postgis-users] Coordinates and SRID - bound overflow, ST_Equals

Hi,

I have a very interesting problem, but I think some of you may have a
solution for that.

When I try to model Russia with EPSG:4326 coordinates, I "run out of
the bounds", e.g.  Russia's extent is so wide, it overlaps the
-180/+180 coordinate range of EPSG:4326, so the most
eastern part is in the western Hemisphere. When inserting points to
the geometry column, PostGIS allows me to insert both (186,45) and
(-174,45) (the former is out of the bounds of the reference, the
latter is okay). When I query the database whether these two points
are equal with ST_Equals, I get a false result , but in reality, they
are the same. I think there should be a version of ST_Equals which
compares the geometries taking into account their SRID.

Thanks,
Balint

-- 
Persics Balint
_______________________________________________
postgis-users mailing list
postgis-users at postgis.refractions.net
http://postgis.refractions.net/mailman/listinfo/postgis-users <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
<http://postgis.refractions.net/mailman/listinfo/postgis-users> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20110120/b2730c1d/attachment.html>