[postgis-devel] ST_Expand unexpected behavior

[David Fuhry]

Oracle spatial returns a four-ordinate rectangle:

SQL> select sdo_aggr_mbr(geom) from v_parcel;

SDO_GEOMETRY(2003, 41100, NULL, SDO_ELEM_INFO_ARRAY(1, 1003, 3), 
SDO_ORDINATE_ARRAY(2210526.17, 645599.584, 2222040.84, 655437.454))

Although the returned geometry's sdo_gtype is that of a 2D polygon, the 
fact that it has only four ordinates means that it should be considered 
to be a 2D bounding box.  This is mentioned in the docs, but the online 
11g doc server is unresponsive, so I can't send a link.

Thanks,

Dave

Obe, Regina wrote:
> Mark,
>> As a matter of interest, can someone with SQL Server 2008 or Oracle 
>> Spatial confirm which datatype is returned for a query of the type 
>> "SELECT ST_Extent(col) FROM table"? Is it a geometry of type POLYGON?
> 
> SQL Server 2008 as I recall doesn't have any aggregate functions
> built-in and I don't see 
> STExtent as an option in the extensions package either.
> 
> http://www.codeplex.com/sqlspatialtools
> 
> 
> I'll be installing SQL Server 2008 RTM probably next week sometime so if
> no one else can verify before then, I'll check to verify.
> 
> I don't have Oracle spatial/locator so can't test that.  I don't see a
> reference to an ST_Extent or Extent function anywhere, so I assume their
> equivalent is
> 
> SDO_AGGR_MBR 
> 
> which sounds to me like a bounding box of some sort.
> 
> http://www.spatialdbadvisor.com/oracle_spatial_tips_tricks/55/oracle-loc
> ator-vs-enterprise-licensing-the-sdogeom-package
> 
> 
> I've cc'd Simon - maybe he can confirm that.
> 
>> Yeah, the current situation with functions returning BOX2D and BOX3D 
>> geometries is not good :(  In SVN trunk I'd like to work on removing 
>> these if possible so that only valid OGC geometries are returned.
> Hmm - Why?  I would presume at a certain point we will start supporting
> 3d geometries more heavily and spatial operators for them and the first
> proto would be the BOX3D.  
> 
> How would your change affect that?
> 
> Thanks,
> Regina
> 
> 
> 
> -----------------------------------------
> 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.
> 
> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel