expr.std

docs/extending.md

@@ -22,6 +22,8 @@ If you want your own library to be recognised too, you can either:
     functions from `narwhals` (e.g. `narwhals.col`, `narwhals.concat`, ...)
   - `Series.__narwhals_series__`: return an object which implements public methods
     from `Narwhals.Series`
+  - `Series.__narwhals_namespace__`: return an object which implements public top-level
+    functions from `narwhals` (e.g. `narwhals.col`, `narwhals.concat`, ...)
 
   If your library doesn't distinguish between lazy and eager, then it's OK for your dataframe
   object to implement both `__narwhals_dataframe__` and `__narwhals_lazyframe__`. In fact,

narwhals/_pandas_like/dataframe.py

@@ -153,7 +153,7 @@ def sort(
         *more_by: str,
         descending: bool | Sequence[bool] = False,
     ) -> Self:
-        flat_keys = flatten([*flatten(by), *more_by])
+        flat_keys = flatten([*flatten([by]), *more_by])
         df = self._dataframe
         if isinstance(descending, bool):
             ascending: bool | list[bool] = not descending

narwhals/_pandas_like/expr.py

@@ -151,6 +151,9 @@ def sum(self) -> Self:
     def mean(self) -> Self:
         return register_expression_call(self, "mean")
 
+    def std(self) -> Self:
+        return register_expression_call(self, "std")
+
     def any(self) -> Self:
         return register_expression_call(self, "any")
 

narwhals/_pandas_like/utils.py

@@ -94,8 +94,12 @@ def maybe_evaluate_expr(df: PandasDataFrame, arg: Any) -> Any:
 def parse_into_exprs(
     implementation: str,
     *exprs: IntoPandasExpr | Iterable[IntoPandasExpr],
+    **named_exprs: IntoPandasExpr,
 ) -> list[PandasExpr]:
-    return [parse_into_expr(implementation, into_expr) for into_expr in flatten(exprs)]
+    out = [parse_into_expr(implementation, into_expr) for into_expr in flatten(exprs)]
+    for name, expr in named_exprs.items():
+        out.append(parse_into_expr(implementation, expr).alias(name))
+    return out
 
 
 def parse_into_expr(implementation: str, into_expr: IntoPandasExpr) -> PandasExpr:

narwhals/expression.py

@@ -169,6 +169,9 @@ def all(self) -> Expr:
     def mean(self) -> Expr:
         return self.__class__(lambda plx: self._call(plx).mean())
 
+    def std(self) -> Expr:
+        return self.__class__(lambda plx: self._call(plx).std())
+
     def sum(self) -> Expr:
         return self.__class__(lambda plx: self._call(plx).sum())
 

tests/test_common.py

@@ -91,6 +91,7 @@ def test_add(df_raw: Any) -> None:
     result = df.with_columns(
         c=nw.col("a") + nw.col("b"),
         d=nw.col("a") - nw.col("a").mean(),
+        e=nw.col("a") - nw.col("a").std(),
     )
     result_native = nw.to_native(result)
     expected = {
@@ -99,6 +100,7 @@ def test_add(df_raw: Any) -> None:
         "z": [7.0, 8.0, 9.0],
         "c": [5, 7, 8],
         "d": [-1.0, 1.0, 0.0],
+        "e": [0.0, 2.0, 1.0],
     }
     compare_dicts(result_native, expected)
 

tests/tpch_q1_test.py

@@ -103,18 +103,16 @@ def test_q1_w_generic_funcs(library: str) -> None:
         )
         .group_by(["l_returnflag", "l_linestatus"])
         .agg(
-            [
-                nw.sum("l_quantity").alias("sum_qty"),
-                nw.sum("l_extendedprice").alias("sum_base_price"),
-                (nw.col("l_extendedprice") * (1 - nw.col("l_discount")))
-                .sum()
-                .alias("sum_disc_price"),
-                nw.col("charge").sum().alias("sum_charge"),
-                nw.mean("l_quantity").alias("avg_qty"),
-                nw.mean("l_extendedprice").alias("avg_price"),
-                nw.mean("l_discount").alias("avg_disc"),
-                nw.len().alias("count_order"),
-            ],
+            nw.sum("l_quantity").alias("sum_qty"),
+            nw.sum("l_extendedprice").alias("sum_base_price"),
+            (nw.col("l_extendedprice") * (1 - nw.col("l_discount")))
+            .sum()
+            .alias("sum_disc_price"),
+            nw.col("charge").sum().alias("sum_charge"),
+            nw.mean("l_quantity").alias("avg_qty"),
+            nw.mean("l_extendedprice").alias("avg_price"),
+            nw.mean("l_discount").alias("avg_disc"),
+            count_order=nw.len(),
         )
         .sort(["l_returnflag", "l_linestatus"])
     )