Hi,


I was also thinking about using some default geometry if no property is given 
but then I had a look on the FES 2.0 standard. I says about BBOX filter, by 
using language which is a bit hard to understand for layman like I:

"If there is only one argument (i.e. one child element specified for the 
operator, the calling service shall apply the operator to the geometric values 
of all the spatial properties of the resource[11]. In this case, the operator 
shall evaluate to true if all tested spatial property values fulfill the 
spatial relationship implied by the operator. Otherwise, the operator shall 
evaluate the specified arguments (i.e. the two child elements) and test whether 
their geometric values satisfy the implied spatial relationship. "

The same applies to DWithin and Beyond operators. For the rest operators FES 
refers to ISO 19125-1:2004, 6.1.14.

So, not one "default geometry" property but "all geometry properties". I think 
I would rather like have that as "any geometry property". If one has something 
like WFS feature type for administrational units which has separate geometries 
for polygon and point geometries, the "all geometries" rule makes BBOX query to 
find the polygon geometry only if also the point geometry happens to be inside 
the box.
Fortunately feature types with multiple geometries are uncommon. But perhaps 
Geoserver could make it better than OGC and accept CQL_FILTER with all these 
variants:
INTERSECTS(the_geom, POLYGON((-90 40, -90 45, -60 45, -60 40, -90 40)))

-          explicit geometry property name
INTERSECTS (all, POLYGON((-90 40, -90 45, -60 45, -60 40, -90 40)))

-          requires that all geometries intersects
INTERSECTS (any, POLYGON((-90 40, -90 45, -60 45, -60 40, -90 40)))

-          it is enough if one geometry intersects
INTERSECTS (POLYGON((-90 40, -90 45, -60 45, -60 40, -90 40))):

-          following what FES 2.0 says about BBOX, Dwithin, and Beyond, means 
the same as "all".

-Jukka Rahkonen-

Andrea Aime wrote:


>Adding the list into the loop

>Cheers
>Andrea
Il 19/ago/2015 21:52, "Andrea Aime" 
<andrea.a...@geo-solutions.it<mailto:andrea.a...@geo-solutions.it>> ha scritto:

Going by memory, GeoServer internally represents the "default geometry"  as an 
empty string, so you may want to try a property name without content, or one 
with an empty CDATA section in place of the name. Let us know if any of that 
works

Cheers
Andrea
Il 19/ago/2015 20:11, "Martin Davis" 
<mtncl...@gmail.com<mailto:mtncl...@gmail.com>> ha scritto:
The WFS spec supports "schemaless" (or perhaps "anonymous" spatial queries 
against featuretypes via the BBOX parameter.  But if a more complex query 
geometry is required (e.g. polygon), it's necessary to create a Filter, which 
then needs to have the spatial propertyname provided.

This means that clients need more metadata about featuretype schemas to do 
polygon queries. This seems a bit inconsistent, since presumably the server 
mechanism can handle schemaless queries (since it can support BBOX).

I suspect I know the answer to this, but I'll ask it anyway...  Is there any 
extension in GeoServer that supports WFS queries with a polygonal query 
geometry WITHOUT requiring a property name?

------------------------------------------------------------------------------

_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net<mailto:Geoserver-users@lists.sourceforge.net>
https://lists.sourceforge.net/lists/listinfo/geoserver-users
------------------------------------------------------------------------------
_______________________________________________
Geoserver-users mailing list
Geoserver-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-users

Reply via email to