[postgis-users] >2D rectangular predicates?

[Chris Nicholas]

well, seems these operators do indeed work for Box3D types, but the 
PostGIS GiST indexing is fundamentally 2D. So building an index on them 
would be interesting, to say the least...

Which in turn makes sense, since  a lot of magic is implied by the 
'Addgeometry_column...' incantation, and the fact that Z coordinates are 
generally not of the same metric as X or Y.

Still, the contrib/cube demo code is most intriguing...

Chris

Chris Nicholas wrote:

> hmmm - perhaps the straight "@", "~", and "&&" will just work with 3D 
> boxes..? that would be way cool...
>
> Chris
>
> Chris Nicholas wrote:
>
>> Hi - I was wondering whether the underlying GiST, etc mechanisms 
>> might support >2D bounding box queries efficiently... I really want 
>> to do fast queries againt BIG tables like :
>>
>> select foo from stuff where CONTAINS( <3/4D bounding box>, the_geom),
>>
>> now, I realize one might do a subselect  in 2D, and then proceed with 
>> measures,  but I'm intrigued by the "cube" stuff I'm seeing in 
>> contrib/cube, and wondering if I might use non-cubic rectangular 
>> regions instead.
>>
>> thanks in advance for any words of wisdom, RTFMs, etc.
>>
>> Chris
>>
>> --------------------------------------------------------------------
>>
>>
>>  [postgis-users] 3d Points
>>
>> *strk at refractions.net * strk at refractions.net 
>> <mailto:postgis-users%40postgis.refractions.net?Subject=%5Bpostgis-users%5D%203d%20Points&In-Reply-To=C8EC7AEB-A485-43B3-98EB-329DFD806870%40geosci.usyd.edu.au> 
>>
>> /Mon Aug 1 11:07:50 PDT 2005/
>>
>>    * Previous message: [postgis-users] 3d Points
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/008836.html> 
>>
>>    * Next message: [postgis-users] 3d Points
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/008858.html> 
>>
>>    * *Messages sorted by:* [ date ]
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/date.html#8838> 
>>
>>      [ thread ]
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/thread.html#8838> 
>>
>>      [ subject ]
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/subject.html#8838> 
>>
>>      [ author ]
>>      
>> <http://postgis.refractions.net/pipermail/postgis-users/2005-August/author.html#8838> 
>>
>>
>>
>> ------------------------------------------------------------------------
>>
>> On Mon, Aug 01, 2005 at 07:47:38PM +1000, Christian Heine wrote:
>> ...
>>
>>> / Is it advisable to use EWKT or is it better to create an extra z- 
>>
>>
>> />/ value column and keep the z-value separated from the lon/lat?
>> /
>> You don't loose anything using the extra Z-value column, you
>> can always extract 2d-version or OGC-standard representation from
>> a 3d object.
>>
>>> / Also,  
>>
>>
>> />/ what is the difference between POINTM and POINT? I couldn't find 
>> any  />/ information explaining the "M".
>> /
>> "M" stand for "Measured". It's an additional coordinate you can use
>> for time or rilevated distance or whatever other parameter you
>> can associate to a vertex of your geometry.
>>
>> In PostGIS we use the M suffix to specify that the
>> tree dimensions given are X,Y,M instead of X,Y,Z.
>>
>> Examples:
>>
>>     POINT(0 0)    <---- X,Y (2D)
>>     POINT(0 0 0)    <---- X,Y,Z (3DZ)
>>     POINTM(0 0 0)    <---- X,Y,M (3DM)
>>     POINT(0 0 0 0)    <---- X,Y,Z,M (4D)
>>
>> Distinction between 3DZ and 3DM is there to store
>> semantic of the third parameter. Some operations
>> would compute an appropriate value for the 3rd
>> dimension if and only if it is a Z (and not an M).
>>
>> --strk;
>>
>> _______________________________________________
>> 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
> .
>