Fix weird summary/by name crash (issue #1255)

datashader/reductions.py

@@ -2171,6 +2171,27 @@ def validate(self, input_dshape):
         for v in self.values:
             v.validate(input_dshape)
 
+        for key, value in zip(self.keys, self.values, strict=True):
+            # Summary keys become Dataset variable names. They cannot collide with
+            # extra dimensions introduced by reductions (e.g. by(cat_column)).
+            dim_names = set()
+            stack = [value]
+            while stack:
+                reduction = stack.pop()
+                if isinstance(reduction, by):
+                    dim_names.add(reduction.cat_column)
+                    stack.append(reduction.reduction)
+                elif isinstance(reduction, where):
+                    stack.append(reduction.selector)
+                elif isinstance(reduction, FloatingNReduction):
+                    dim_names.add("n")
+
+            if key in dim_names:
+                raise ValueError(
+                    f"Invalid summary reduction name {key!r}: it conflicts with "
+                    f"a generated dimension name. Rename the summary key or reduction."
+                )
+
         # Check that any included FloatingNReductions have the same n values.
         n_values = []
         for v in self.values:

datashader/tests/test_xarray.py

@@ -2,6 +2,7 @@
 from copy import deepcopy
 import numpy as np
 from numpy import nan
+import pandas as pd
 import xarray as xr
 
 import datashader as ds
@@ -47,6 +48,18 @@ def assert_eq(agg, b):
     assert agg.equals(b)
 
 
+def test_summary_name_conflicts_with_by_dimension():
+    source = xr.Dataset(coords={
+        "x": xr.DataArray(np.array([0, 0, 1, 1]), dims="record"),
+        "y": xr.DataArray(np.array([0, 1, 0, 1]), dims="record"),
+        "foo foo": xr.DataArray(pd.Categorical(["a", "a", "b", "b"]), dims="record"),
+    })
+    agg = ds.summary(**{"foo foo": ds.by("foo foo")})
+    msg = "Invalid summary reduction name 'foo foo': it conflicts with a generated dimension name"
+    with pytest.raises(ValueError, match=msg):
+        c.points(source, "x", "y", agg)
+
+
 @pytest.mark.parametrize("source", [xda, xdda, xds, xdds])
 def test_count(source):
     if source is None: