Hi, is it possible to have spatial selection (geof:sfIntersects) against a WKT literal?
By using the equivalent geof:relate(?g1, ?g2, geo:intersects) it is clear that ?g2 must be a geometry.
Obviously one way would be to create (insert) a few triples with an imaginary geometry which would be linked to the WKT literal, but that is not satisfactory in general.
not sure if I understand correctly, but in GeoSPARQL
you either use the relation as a property function (e.g. geo:sfIntersects) which then expects a geometry or a feature (there are rewriting rules in the GeoSPARQL standard, in the end it's nothing more then convenience and sometimes also needs inferencing enabled to infer that something is a geo:Feature or a geo:Geometry)
or you use the filter expression (e.g. geof:sfIntersects) which works only the serialization literal level and expects literals
Topology Vocabulary Extension
geo:sfIntersects with geo:SpatialObject as domain and range.
Stardog does not seem to support either version of the GeoSPARQL standard (geof:sfIntersects, geo:sfIntersects).
The Query Rewrite Extension defines a set of RIF rules [15] that use topological
extension functions defined in Geometry Topology Extension (Ch.9) to establish the existence of direct topological predicates defined in Topology Vocabulary Extension (Ch. 7).
Now my question is this. Is there any way I express a spatial selection query, like the one you included in your answer, in Stardog:
Also you can check in my other post/question " [Geof:relate boolean function returns incosistent results compared to the predicate form (BGP)]" for a sample dataset with inconsistencies between the two forms of the geof:relate function/predicate.