Fixed bug in canonicalizing xarray data

holoviews/core/data/grid.py

@@ -154,16 +154,16 @@ def canonicalize(cls, dataset, data, coord_dims=None):
         dims = [name for name in coord_dims[::-1]
                 if isinstance(cls.coords(dataset, name), np.ndarray)]
         dropped = [dims.index(d) for d in dims if d not in dataset.kdims]
-        inds = [dims.index(kd.name) for kd in dataset.kdims]
-        inds += dropped
+        inds = [dims.index(kd.name)for kd in dataset.kdims]
+        inds = [i - sum([1 for d in dropped if i>=d]) for i in inds]
+        if dropped:
+            data = data.squeeze(axis=tuple(dropped))
         if inds:
             data = data.transpose(inds)
 
         # Allow lower dimensional views into data
         if len(dataset.kdims) < 2:
             data = data.flatten()
-        elif dropped:
-            data = data.squeeze(axis=tuple(range(len(dropped))))
         return data
 
 

holoviews/core/data/xarray.py

@@ -50,7 +50,8 @@ def init(cls, eltype, data, kdims, vdims):
             elif len(vdim_param.default) == 1:
                 vdim = vdim_param.default[0]
             vdims = [vdim]
-            kdims = [Dimension(d) for d in data.dims[::-1]]
+            if not kdims:
+                kdims = [Dimension(d) for d in data.dims[::-1]]
             data = data.to_dataset(name=vdim.name)
         elif not isinstance(data, xr.Dataset):
             if kdims is None:
@@ -148,7 +149,6 @@ def groupby(cls, dataset, dimensions, container_type, group_type, **kwargs):
 
     @classmethod
     def coords(cls, dataset, dim, ordered=False, expanded=False):
-        dim = dataset.get_dimension(dim, strict=True).name
         if expanded:
             return util.expand_grid_coords(dataset, dim)
         data = np.atleast_1d(dataset.data[dim].data)
@@ -162,7 +162,7 @@ def values(cls, dataset, dim, expanded=True, flat=True):
         dim = dataset.get_dimension(dim, strict=True)
         data = dataset.data[dim.name].data
         if dim in dataset.vdims:
-            coord_dims = dataset.data[dim.name].dims
+            coord_dims = list(dataset.data.dims.keys())[::-1]
             if dask and isinstance(data, dask.array.Array):
                 data = data.compute()
             data = cls.canonicalize(dataset, data, coord_dims=coord_dims)

tests/testdataset.py

@@ -1055,6 +1055,15 @@ def setUp(self):
         self.init_column_data()
         self.init_grid_data()
 
+    def test_xarray_dataset_with_scalar_dim_canonicalize(self):
+        import xarray as xr
+        xs = [0, 1]
+        ys = [0.1, 0.2, 0.3]
+        zs = np.array([[[0, 1], [2, 3], [4, 5]]])
+        xrarr = xr.DataArray(zs, coords={'x': xs, 'y': ys, 't': [1]}, dims=['t', 'y', 'x'])
+        ds = Dataset(xrarr, kdims=['x', 'y'], vdims=['z'], datatype=['xarray'])
+        self.assertEqual(ds.dimension_values(2, flat=False).ndim, 2)
+
     # Disabled tests for NotImplemented methods
     def test_dataset_add_dimensions_values_hm(self):
         raise SkipTest("Not supported")