v.select: exact meaning of disjoint operator

[Moritz Lennert]

What exactly does the disjoint operator do in the GEOS operators of
v.select (6.5+) ? It is described as

"disjoint - features do not spatially intersect"

Which sounds pretty much like the reverse of the intersect operator.

However, when I try the following in the nc_spm_06 NC data set:

v.extract input=roadsmajor@PERMANENT output=cary list=1004
v.select -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
btype=area output=roads_disjoint_cary operator=disjoint

I get:

http://geog-pc40.ulb.ac.be/grass/misc/disjoint.png

whereas

v.select -r -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
btype=area output=roads_reverse_intersect_cary operator=intersects

gives me:

http://geog-pc40.ulb.ac.be/grass/misc/reverse_intersect.png

Is this a bug in the disjoint operator, or am I misinterpreting its
meaning ?

Moritz


_______________________________________________
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user

[Martin Landa]

Hi,

2009/6/26 Moritz Lennert <[hidden email]>:

> "disjoint - features do not spatially intersect"
>
> Which sounds pretty much like the reverse of the intersect operator.
>
> However, when I try the following in the nc_spm_06 NC data set:
>
> v.extract input=roadsmajor@PERMANENT output=cary list=1004
> v.select -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
> btype=area output=roads_disjoint_cary operator=disjoint
>
> I get:
>
> http://geog-pc40.ulb.ac.be/grass/misc/disjoint.png
>
> whereas
>
> v.select -r -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
> btype=area output=roads_reverse_intersect_cary operator=intersects
>
> gives me:
>
> http://geog-pc40.ulb.ac.be/grass/misc/reverse_intersect.png
>
> Is this a bug in the disjoint operator, or am I misinterpreting its meaning

It's a bug in v.select. Fixed in r38250. Currently `ope=disjoint` =
`-r operator=intersect`. It would be better to change v.select's
algorithm to select features, and to remove this ugly hack.

Martin

--
Martin Landa <landa.martin gmail.com> * http://gama.fsv.cvut.cz/~landa
_______________________________________________
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user

[Moritz Lennert]

On 05/07/09 15:13, Martin Landa wrote:

> Hi,
>
> 2009/6/26 Moritz Lennert <[hidden email]>:
>> "disjoint - features do not spatially intersect"
>>
>> Which sounds pretty much like the reverse of the intersect operator.
>>
>> However, when I try the following in the nc_spm_06 NC data set:
>>
>> v.extract input=roadsmajor@PERMANENT output=cary list=1004
>> v.select -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
>> btype=area output=roads_disjoint_cary operator=disjoint
>>
>> I get:
>>
>> http://geog-pc40.ulb.ac.be/grass/misc/disjoint.png
>>
>> whereas
>>
>> v.select -r -g ainput=roadsmajor@PERMANENT atype=line binput=cary@user1
>> btype=area output=roads_reverse_intersect_cary operator=intersects
>>
>> gives me:
>>
>> http://geog-pc40.ulb.ac.be/grass/misc/reverse_intersect.png
>>
>> Is this a bug in the disjoint operator, or am I misinterpreting its meaning
>
> It's a bug in v.select. Fixed in r38250. Currently `ope=disjoint` =
> `-r operator=intersect`. It would be better to change v.select's
> algorithm to select features, and to remove this ugly hack.

Ok, thanks for following up on this. Should we file a bug report for the
selection algorithm ? Could you give me more details on the problems
with it ?

Moritz
_______________________________________________
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user

[Martin Landa]

Hi,

2009/7/7 Moritz Lennert <[hidden email]>:
> Ok, thanks for following up on this. Should we file a bug report for the
> selection algorithm ? Could you give me more details on the problems with it

the problem is that v.select was originally designed only for
'overlaps' operator. So only features with MBR from B which overlaps
MBR of feature from A are tested. Which is not the case of disjoint
operator.

Martin

--
Martin Landa <landa.martin gmail.com> * http://gama.fsv.cvut.cz/~landa
_______________________________________________
grass-user mailing list
[hidden email]
http://lists.osgeo.org/mailman/listinfo/grass-user