unxfail for dask, but is it a good idea?

narwhals/_dask/namespace.py

@@ -24,12 +24,12 @@
 from narwhals._expression_parsing import combine_evaluate_output_names
 from narwhals.typing import CompliantNamespace
 
-if TYPE_CHECKING:
-    try:
-        import dask.dataframe.dask_expr as dx
-    except ModuleNotFoundError:
-        import dask_expr as dx
+try:
+    import dask.dataframe.dask_expr as dx
+except ModuleNotFoundError:
+    import dask_expr as dx
 
+if TYPE_CHECKING:
     from typing_extensions import Self
 
     from narwhals.dtypes import DType
@@ -157,7 +157,11 @@ def func(df: DaskLazyFrame) -> list[dx.Series]:
 
     def sum_horizontal(self: Self, *exprs: DaskExpr) -> DaskExpr:
         def func(df: DaskLazyFrame) -> list[dx.Series]:
-            series = [s.fillna(0) for _expr in exprs for s in _expr(df)]
+            series = [
+                s.fillna(0) if isinstance(s, dx.Series) else 0 if s is None else s
+                for _expr in exprs
+                for s in _expr(df)
+            ]
             return [reduce(operator.add, series)]
 
         return DaskExpr(

tests/expr_and_series/sum_horizontal_test.py

@@ -47,11 +47,7 @@ def test_sumh_all(constructor: Constructor) -> None:
     assert_equal_data(result, expected)
 
 
-def test_sumh_aggregations(
-    constructor: Constructor, request: pytest.FixtureRequest
-) -> None:
-    if "dask" in str(constructor):
-        request.applymarker(pytest.mark.xfail)
+def test_sumh_aggregations(constructor: Constructor) -> None:
     data = {"a": [1, 2, 3], "b": [10, 20, 30]}
     df = nw.from_native(constructor(data))
     result = df.select(nw.sum_horizontal(nw.all().mean().name.suffix("_foo")))
@@ -64,7 +60,7 @@ def test_sumh_aggregations(
 def test_sumh_transformations(
     constructor: Constructor, request: pytest.FixtureRequest
 ) -> None:
-    if any(x in str(constructor) for x in ("dask", "duckdb")):
+    if any(x in str(constructor) for x in ("duckdb",)):
         request.applymarker(pytest.mark.xfail)
     data = {"a": [1, 2, 3], "b": [4, 5, 6], "c": [7, 8, 9]}
     df = nw.from_native(constructor(data))