BUG: Fix pyarrow categoricals not working for pivot and multiindex (#61193)

* BUG: Fix bug with DataFrame.pivot and .set_index not compatible with pyarrow dictionary categoricals

Relates to #53051

Code for fix taken and adapted from #59099

* TST: Add tests for faulty behavior relating to pyarrow categoricals

* CLN: Fix issues reported by pre-commit hooks

* TST: Fix failing tests for minimum version by ignoring obsolete deprecation warning

* DOC: Add entry for bugfix to whatsnew v3.0.0

* CLN: Refactor code and clean up according to PR feedback

* CLN: Refactor code and clean up according to PR feedback

* CLN: Refactor tests to adress PR feedback

* CLN: Refactor tests to adress PR feedback

Author: robin-mader-bis

Diff for: doc/source/whatsnew/v3.0.0.rst

@@ -639,6 +639,7 @@ Bug fixes
 Categorical
 ^^^^^^^^^^^
 - Bug in :func:`Series.apply` where ``nan`` was ignored for :class:`CategoricalDtype` (:issue:`59938`)
+- Bug in :meth:`DataFrame.pivot` and :meth:`DataFrame.set_index` raising an ``ArrowNotImplementedError`` for columns with pyarrow dictionary dtype (:issue:`53051`)
 - Bug in :meth:`Series.convert_dtypes` with ``dtype_backend="pyarrow"`` where empty :class:`CategoricalDtype` :class:`Series` raised an error or got converted to ``null[pyarrow]`` (:issue:`59934`)
 -
 

Diff for: pandas/core/arrays/categorical.py

@@ -452,7 +452,7 @@ def __init__(
                 if isinstance(values, Index):
                     arr = values._data._pa_array.combine_chunks()
                 else:
-                    arr = values._pa_array.combine_chunks()
+                    arr = extract_array(values)._pa_array.combine_chunks()
                 categories = arr.dictionary.to_pandas(types_mapper=ArrowDtype)
                 codes = arr.indices.to_numpy()
                 dtype = CategoricalDtype(categories, values.dtype.pyarrow_dtype.ordered)

Diff for: pandas/tests/reshape/test_pivot.py

@@ -15,6 +15,7 @@
 
 import pandas as pd
 from pandas import (
+    ArrowDtype,
     Categorical,
     DataFrame,
     Grouper,
@@ -2851,3 +2852,31 @@ def test_pivot_margins_with_none_index(self):
             ),
         )
         tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.filterwarnings("ignore:Passing a BlockManager:DeprecationWarning")
+    def test_pivot_with_pyarrow_categorical(self):
+        # GH#53051
+        pa = pytest.importorskip("pyarrow")
+
+        df = DataFrame(
+            {"string_column": ["A", "B", "C"], "number_column": [1, 2, 3]}
+        ).astype(
+            {
+                "string_column": ArrowDtype(pa.dictionary(pa.int32(), pa.string())),
+                "number_column": "float[pyarrow]",
+            }
+        )
+
+        df = df.pivot(columns=["string_column"], values=["number_column"])
+
+        multi_index = MultiIndex.from_arrays(
+            [["number_column", "number_column", "number_column"], ["A", "B", "C"]],
+            names=(None, "string_column"),
+        )
+        df_expected = DataFrame(
+            [[1.0, np.nan, np.nan], [np.nan, 2.0, np.nan], [np.nan, np.nan, 3.0]],
+            columns=multi_index,
+        )
+        tm.assert_frame_equal(
+            df, df_expected, check_dtype=False, check_column_type=False
+        )

Diff for: pandas/tests/test_multilevel.py

@@ -5,6 +5,7 @@
 
 import pandas as pd
 from pandas import (
+    ArrowDtype,
     DataFrame,
     MultiIndex,
     Series,
@@ -318,6 +319,34 @@ def test_multiindex_dt_with_nan(self):
         expected = Series(["a", "b", "c", "d"], name=("sub", np.nan))
         tm.assert_series_equal(result, expected)
 
+    @pytest.mark.filterwarnings("ignore:Passing a BlockManager:DeprecationWarning")
+    def test_multiindex_with_pyarrow_categorical(self):
+        # GH#53051
+        pa = pytest.importorskip("pyarrow")
+
+        df = DataFrame(
+            {"string_column": ["A", "B", "C"], "number_column": [1, 2, 3]}
+        ).astype(
+            {
+                "string_column": ArrowDtype(pa.dictionary(pa.int32(), pa.string())),
+                "number_column": "float[pyarrow]",
+            }
+        )
+
+        df = df.set_index(["string_column", "number_column"])
+
+        df_expected = DataFrame(
+            index=MultiIndex.from_arrays(
+                [["A", "B", "C"], [1, 2, 3]], names=["string_column", "number_column"]
+            )
+        )
+        tm.assert_frame_equal(
+            df,
+            df_expected,
+            check_index_type=False,
+            check_column_type=False,
+        )
+
 
 class TestSorted:
     """everything you wanted to test about sorting"""