GH-35250: [Python] Add test for datetime column conversion to pandas (#…

…35546)

### Rationale for this change
In pandas version 2.1.0 `__from_arrow__` method has been implemented for `DatetimeTZDtype` extension type. See pandas-dev/pandas#52201.

Due to this change, the conversion from pyarrow table column with datetime uses the newly implemented `__from_arrow__` method for the pandas dtype and `pandas_api.series` to construct the pandas series. We had to pass the name in the `pandas_api.series` also, to fix failing tests in #35248.

This PR adds explicit test for this change.
* Closes: #35250

Authored-by: Alenka Frim <frim.alenka@gmail.com>
Signed-off-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>

python/pyarrow/tests/test_pandas.py

@@ -4816,3 +4816,17 @@ def test_unhashable_map_keys_with_pydicts():
         for tup1, tup2 in zip(row1, row2):
             assert np.array_equal(tup1[0], tup2[0])
             assert tup1[1] == tup2[1]
+
+
+def test_column_conversion_for_datetime():
+    # GH-35235
+    # pandas implemented __from_arrow__ for DatetimeTZDtype
+    # https://github.com/pandas-dev/pandas/pull/52201
+    arr = pd.Series(pd.date_range("2012", periods=2, tz="Europe/Brussels"),
+                    name="datetime_column")
+    table = pa.table({"datetime_column": pa.array(arr)})
+    table_col = table.column("datetime_column")
+
+    result = table_col.to_pandas()
+    assert result.name == "datetime_column"
+    tm.assert_series_equal(result, arr)