[postgis-users] ST_Contains - WGS84
robe.dnd at cityofboston.gov
Fri Jun 27 08:37:22 PDT 2008
I don't think the relation functions take into consideration projection at all. They just work against the raw coordinates you give them. This is why you can have SRID = -1 and things would be fine anyway.
As I understand it, they just perform operations on the coordinates given to them and verify that the two geometries are in the same coordinate system otherwise the analysis would be pointless since you would be comparing apples and oranges.
Now I'm not quite sure what happens in the wrap around case - if you have a geometry that crosses around say between 0 and 360 degrees (with SRID 4326). In a spherical world you would see they are close but since PostGIS has no concept of geodetic space when looking at it in a pure cartesian system (where the sphere is rolled out as is the really the case in 4326) they would be on distant ends of the map. But I suppose you can then shift the degrees or something to compensate.
From: postgis-users-bounces at postgis.refractions.net [mailto:postgis-users-bounces at postgis.refractions.net] On Behalf Of postgis.domain.thewild at spamgourmet.com
Sent: Friday, June 27, 2008 10:46 AM
To: postgis-users at postgis.refractions.net
Subject: Re: [postgis-users] ST_Contains - WGS84
+postgis+thewild+172cd4478a.postgis.domain.thewild#spamgourmet.com at spamgourmet.com
a écrit :
> Markus Schaber - schabi at logix-tt.com a écrit :
>> I'm not really sure about that, it may be that the rubberbands get
>> mangled into curves in certain projections. And so vertexes from other
>> polygons might "jump" on the other side of the rubberband.
> Indeed, I would assume that too.
> Projections do not preserve parallelism, so I guess that the same
> geometries projected in two different srids could have different
> ST_Relate results.
> If you want "true results", I think you have to go for a real coordinate
> space like 4326.
Thinking a bit more about it, maybe those spatial relation checks are
always internally performed in unprojected coordinates.
That would be logical, since projections are just a representation of
real things, relations should be checked on the real things rather than
on their representations.
postgis-users mailing list
postgis-users at postgis.refractions.net
The substance of this message, including any attachments, may be
confidential, legally privileged and/or exempt from disclosure
pursuant to Massachusetts law. It is intended
solely for the addressee. If you received this in error, please
contact the sender and delete the material from any computer.
More information about the postgis-users