perf: avoid redundant copies when resetting index for pandas

narwhals/_pandas_like/group_by.py

@@ -11,6 +11,7 @@
 from narwhals._expression_parsing import is_simple_aggregation
 from narwhals._expression_parsing import parse_into_exprs
 from narwhals._pandas_like.utils import native_series_from_iterable
+from narwhals._pandas_like.utils import reset_index_no_copy
 from narwhals.utils import Implementation
 from narwhals.utils import remove_prefix
 
@@ -187,13 +188,14 @@ def agg_pandas(  # noqa: PLR0915
             ]
             result_simple_aggs = result_simple_aggs.rename(
                 columns=name_mapping, copy=False
-            ).reset_index()
+            )
+            reset_index_no_copy(result_simple_aggs, native_namespace)
         if nunique_aggs:
             result_nunique_aggs = grouped[list(nunique_aggs.values())].nunique(
                 dropna=False
             )
             result_nunique_aggs.columns = list(nunique_aggs.keys())
-            result_nunique_aggs = result_nunique_aggs.reset_index()
+            reset_index_no_copy(result_nunique_aggs, native_namespace)
         if simple_aggs and nunique_aggs:
             if (
                 set(result_simple_aggs.columns)
@@ -259,6 +261,6 @@ def func(df: Any) -> Any:
     else:  # pragma: no cover
         result_complex = grouped.apply(func)
 
-    result = result_complex.reset_index()
+    reset_index_no_copy(result_complex, native_namespace)
 
-    return from_dataframe(result.loc[:, output_names])
+    return from_dataframe(result_complex.loc[:, output_names])

narwhals/_pandas_like/series.py

@@ -14,6 +14,7 @@
 from narwhals._pandas_like.utils import narwhals_to_native_dtype
 from narwhals._pandas_like.utils import native_series_from_iterable
 from narwhals._pandas_like.utils import native_to_narwhals_dtype
+from narwhals._pandas_like.utils import reset_index_no_copy
 from narwhals._pandas_like.utils import set_axis
 from narwhals._pandas_like.utils import to_datetime
 from narwhals._pandas_like.utils import validate_column_comparand
@@ -598,7 +599,8 @@ def value_counts(
             dropna=False,
             sort=False,
             normalize=normalize,
-        ).reset_index()
+        )
+        reset_index_no_copy(val_count, self.__native_namespace__())
 
         val_count.columns = [index_name_, value_name_]
 

narwhals/_pandas_like/utils.py

@@ -590,3 +590,16 @@ def calculate_timestamp_date(s: pd.Series, time_unit: str) -> pd.Series:
     else:
         result = s * 1_000
     return result
+
+
+def reset_index_no_copy(native_series_or_frame: Any, native_namespace: Any) -> None:
+    if set(native_series_or_frame.index.names).intersection(
+        native_series_or_frame.columns
+    ):
+        msg = f"Cannot insert column with name {native_series_or_frame.index.name} into dataframe with columns {native_series_or_frame.columns}"
+        raise ValueError(msg)
+    for i, name in enumerate(native_series_or_frame.index.names):
+        native_series_or_frame[name] = native_series_or_frame.index.get_level_values(i)
+    native_series_or_frame.index = native_namespace.RangeIndex(
+        len(native_series_or_frame)
+    )