[postgis-devel] POINT EMPTY in WKB

[Paul Ramsey]

On Mon, Dec 6, 2010 at 1:00 PM, Paul Ramsey <pramsey at opengeo.org> wrote:
> It's just POINT EMPTY. Everything else has a representation. In fact,
> I think Mats checked for me and SQLServer uses the same semantic as I
> am for empties in WKB, npoints=0 or ngeoms=0 or nrings=0 depending on
> the type being represented.

http://mateusz.loskot.net/2010/02/26/sqlgeometry-and-point-empty-in-wkb/

Mats did, and the answer is SQLServer turns 'POINT EMPTY' into
'MULTIPOINT EMPTY'

P

> And I think for POINT EMPTY they transform to GEOMETRYCOLLECTION
> EMPTY. (which is what we do now, see
>
>  SELECT AsText('POINT EMPTY'::geometry);
>
> P.
>
> On Mon, Dec 6, 2010 at 12:48 PM, Paragon Corporation <lr at pcorp.us> wrote:
>>
>>
>>> A final possibility would be to hack POINT EMPTY into WKB, as
>>> something like POINT(nan nan nan) or POINT(inf inf inf) or POINT(-0 -0
>>> -0). There would be a danger there in platforms that have different
>>> interpretations of what nan, inf or -0 are in binary.
>>
>> I the only issue with POINT EMPTY or ..anyspecifictype EMPTY
>>
>> I'm a bit bothered about the dump issue you describe -- not being able to
>> restore exactly what you thought you had.
>>
>> Thanks,
>> Regina
>>
>>
>>
>> _______________________________________________
>> postgis-devel mailing list
>> postgis-devel at postgis.refractions.net
>> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>>
>