feat(python)!: Remove Groupby.pivot

py-polars/docs/source/reference/dataframe/groupby.rst

@@ -20,7 +20,6 @@ This namespace is available after calling :code:`DataFrame.groupby(...)`.
     GroupBy.median
     GroupBy.min
     GroupBy.n_unique
-    GroupBy.pivot
     GroupBy.quantile
     GroupBy.sum
     GroupBy.tail

py-polars/docs/source/reference/dataframe/index.rst

@@ -16,7 +16,6 @@ This page gives an overview of all public DataFrame methods.
    groupby
    modify_select
    miscellaneous
-   pivot
 
 .. currentmodule:: polars
 

py-polars/polars/internals/dataframe/groupby.py

@@ -5,7 +5,6 @@
 from typing import TYPE_CHECKING, Callable, Generic, Iterator, Sequence, TypeVar
 
 import polars.internals as pli
-from polars.internals.dataframe.pivot import PivotOps
 from polars.utils import _timedelta_to_pl_duration, is_str_sequence
 
 if TYPE_CHECKING:
@@ -361,71 +360,6 @@ def tail(self, n: int = 5) -> DF:
         )
         return self._dataframe_class._from_pydf(df._df)
 
-    def pivot(
-        self, pivot_column: str | list[str], values_column: str | list[str]
-    ) -> PivotOps[DF]:
-        """
-        Do a pivot operation.
-
-        The pivot operation is based on the group key, a pivot column and an aggregation
-        function on the values column.
-
-        .. deprecated:: 0.13.23
-            `DataFrame.groupby.pivot` will be removed in favour of `DataFrame.pivot`.
-
-        Parameters
-        ----------
-        pivot_column
-            Column to pivot.
-        values_column
-            Column that will be aggregated.
-
-        Notes
-        -----
-        Polars'/arrow memory is not ideal for transposing operations like pivots.
-        If you have a relatively large table, consider using a groupby over a pivot.
-
-        Examples
-        --------
-        >>> df = pl.DataFrame(
-        ...     {
-        ...         "foo": ["one", "one", "one", "two", "two", "two"],
-        ...         "bar": ["A", "B", "C", "A", "B", "C"],
-        ...         "baz": [1, 2, 3, 4, 5, 6],
-        ...     }
-        ... )
-        >>> df.groupby("foo", maintain_order=True).pivot(  # doctest: +SKIP
-        ...     pivot_column="bar", values_column="baz"
-        ... ).first()
-        shape: (2, 4)
-        ┌─────┬─────┬─────┬─────┐
-        │ foo ┆ A   ┆ B   ┆ C   │
-        │ --- ┆ --- ┆ --- ┆ --- │
-        │ str ┆ i64 ┆ i64 ┆ i64 │
-        ╞═════╪═════╪═════╪═════╡
-        │ one ┆ 1   ┆ 2   ┆ 3   │
-        │ two ┆ 4   ┆ 5   ┆ 6   │
-        └─────┴─────┴─────┴─────┘
-
-        """
-        warnings.warn(
-            "`DataFrame.groupby.pivot` is deprecated and will be removed in a future"
-            " version. Use `DataFrame.pivot` instead.",
-            DeprecationWarning,
-            stacklevel=2,
-        )
-        if isinstance(pivot_column, str):
-            pivot_column = [pivot_column]
-        if isinstance(values_column, str):
-            values_column = [values_column]
-        return PivotOps(
-            self._df,
-            self.by,  # type: ignore[arg-type]
-            pivot_column,
-            values_column,
-            dataframe_class=self._dataframe_class,
-        )
-
     def first(self) -> pli.DataFrame:
         """
         Aggregate the first values in the group.

py-polars/tests/unit/test_df.py

@@ -2214,13 +2214,6 @@ class SubClassedDataFrame(pl.DataFrame):
     groupby = subclassed_df.groupby("a")
     assert isinstance(groupby.agg(pl.count()), SubClassedDataFrame)
 
-    # Round-trips to PivotOps and back should also preserve subclass
-    with pytest.deprecated_call():
-        assert isinstance(
-            groupby.pivot(pivot_column="a", values_column="b").first(),
-            SubClassedDataFrame,
-        )
-
 
 def test_explode_empty() -> None:
     df = (

py-polars/tests/unit/test_pivot.py

@@ -73,34 +73,6 @@ def test_pivot_aggregate(agg_fn: PivotAgg, expected_rows: list[tuple[Any]]) -> N
     assert result.rows() == expected_rows
 
 
-@pytest.mark.parametrize(
-    ("agg_fn", "expected_rows"),
-    [
-        ("first", [("a", 2, None, None), ("b", None, None, 10)]),
-        ("count", [("a", 2, None, None), ("b", None, 2, 1)]),
-        ("min", [("a", 2, None, None), ("b", None, 8, 10)]),
-        ("max", [("a", 4, None, None), ("b", None, 8, 10)]),
-        ("sum", [("a", 6, None, None), ("b", None, 8, 10)]),
-        ("mean", [("a", 3.0, None, None), ("b", None, 8.0, 10.0)]),
-        ("median", [("a", 3.0, None, None), ("b", None, 8.0, 10.0)]),
-    ],
-)
-def test_pivot_groupby_aggregate(
-    agg_fn: PivotAgg, expected_rows: list[tuple[Any]]
-) -> None:
-    df = pl.DataFrame(
-        {
-            "a": [1, 1, 2, 2, 3],
-            "b": ["a", "a", "b", "b", "b"],
-            "c": [2, 4, None, 8, 10],
-        }
-    )
-    with pytest.deprecated_call():
-        pivot = df.groupby("b").pivot(pivot_column="a", values_column="c")
-    result = getattr(pivot, agg_fn)()
-    assert result.rows() == expected_rows
-
-
 def test_pivot_categorical_3968() -> None:
     df = pl.DataFrame(
         {