Resolve custom-labelled dimensions by name in GeoPandasInterface

geoviews/data/geopandas.py

@@ -13,7 +13,7 @@
 )
 from holoviews.core.data.interface import DataError
 from holoviews.core.data.spatialpandas import get_value_array
-from holoviews.core.dimension import dimension_name
+from holoviews.core.dimension import Dimension, dimension_name
 from holoviews.core.util import isscalar, unique_array, unique_iterator
 from holoviews.element import Path
 
@@ -257,24 +257,39 @@ def dimension_type(cls, dataset, dim):
         else:
             return float
 
+    @classmethod
+    def _resolve_dimension(cls, dataset, dim):
+        """
+        Resolves a dimension spec to a Dimension of the dataset.
+
+        A Dimension built from a string spec (e.g. by a ``dim('name')``
+        expression) carries the name as its label, and Dimension equality
+        compares labels, so it would not match a dimension with a custom
+        label. Fall back to looking up the name, which matches both names
+        and labels.
+        """
+        resolved = dataset.get_dimension(dim)
+        if resolved is None and isinstance(dim, Dimension):
+            resolved = dataset.get_dimension(dim.name)
+        return resolved
+
     @classmethod
     def isscalar(cls, dataset, dim, per_geom=False):
         """
         Tests if dimension is scalar in each subpath.
         """
-        dim = dataset.get_dimension(dim)
+        dim = cls._resolve_dimension(dataset, dim)
         geom_dims = cls.geom_dims(dataset)
         if dim in geom_dims:
             return False
         elif per_geom:
             return all(isscalar(v) or len(list(unique_array(v))) == 1
                        for v in dataset.data[dim.name])
-        dim = dataset.get_dimension(dim)
         return len(dataset.data[dim.name].unique()) == 1
 
     @classmethod
     def range(cls, dataset, dim):
-        dim = dataset.get_dimension(dim)
+        dim = cls._resolve_dimension(dataset, dim)
         geom_dims = cls.geom_dims(dataset)
         if dim in geom_dims:
             col = cls.geo_column(dataset.data)
@@ -356,7 +371,7 @@ def redim(cls, dataset, dimensions):
 
     @classmethod
     def values(cls, dataset, dimension, expanded=True, flat=True, compute=True, keep_index=False):
-        dimension = dataset.get_dimension(dimension)
+        dimension = cls._resolve_dimension(dataset, dimension)
         geom_dims = dataset.interface.geom_dims(dataset)
         data = dataset.data
         isgeom = (dimension in geom_dims)

geoviews/tests/data/test_geopandas.py

@@ -212,3 +212,23 @@ def test_geopandas_dataframe_with_different_dtype_column(self):
         )
         gdf = geopandas.GeoDataFrame(df, geometry=geopandas.points_from_xy(df.x, df.y))
         render(Points(gdf))
+
+    def test_geopandas_vdim_with_custom_label_resolved_by_name(self):
+        # Fix for https://github.com/holoviz/geoviews/issues/840
+        # A dim() expression carries a Dimension whose label is the column
+        # name, which must still resolve against a vdim with a custom label.
+        from holoviews import dim
+        gdf = geopandas.GeoDataFrame(
+            {"status": ["safe", "endangered", "safe"]},
+            geometry=geopandas.points_from_xy([0, 1, 2], [0, 1, 2]),
+        )
+        points = Points(gdf, vdims=[("status", "Conservation Status")])
+        probe = dim("status").dimension
+        assert self.interface.isscalar(points, probe, per_geom=True)
+        assert_data_equal(
+            self.interface.values(points, probe),
+            np.array(["safe", "endangered", "safe"], dtype=object),
+        )
+        assert self.interface.range(points, probe) == ("endangered", "safe")
+        color = dim("status").categorize({"safe": "green", "endangered": "red"})
+        render(points.opts(color=color))