Backport PR #47347 on branch 1.4.x (REGR: Regression in to_csv for ea…

… dtype categorical) (#47388)

* Backport PR #47347: REGR: Regression in to_csv for ea dtype categorical

* inclusive -> closed

Co-authored-by: Patrick Hoefler <61934744+phofl@users.noreply.github.com>
Co-authored-by: Simon Hawkins <simonjayhawkins@gmail.com>

doc/source/whatsnew/v1.4.3.rst

@@ -15,6 +15,7 @@ including other versions of pandas.
 Fixed regressions
 ~~~~~~~~~~~~~~~~~
 - Fixed regression in :meth:`DataFrame.replace` when the replacement value was explicitly ``None`` when passed in a dictionary to ``to_replace`` also casting other columns to object dtype even when there were no values to replace (:issue:`46634`)
+- Fixed regression in :meth:`DataFrame.to_csv` raising error when :class:`DataFrame` contains extension dtype categorical column (:issue:`46297`, :issue:`46812`)
 - Fixed regression in representation of ``dtypes`` attribute of :class:`MultiIndex` (:issue:`46900`)
 - Fixed regression when setting values with :meth:`DataFrame.loc` updating :class:`RangeIndex` when index was set as new column and column was updated afterwards (:issue:`47128`)
 - Fixed regression in :meth:`DataFrame.nsmallest` led to wrong results when ``np.nan`` in the sorting column (:issue:`46589`)

pandas/core/internals/blocks.py

@@ -2160,7 +2160,7 @@ def to_native_types(
     **kwargs,
 ) -> np.ndarray:
     """convert to our native types format"""
-    if isinstance(values, Categorical):
+    if isinstance(values, Categorical) and values.categories.dtype.kind in "Mm":
         # GH#40754 Convert categorical datetimes to datetime array
         values = take_nd(
             values.categories._values,

pandas/tests/frame/methods/test_to_csv.py

@@ -1333,3 +1333,32 @@ def test_to_csv_na_quoting(self):
         )
         expected = '""\n""\n'
         assert result == expected
+
+    def test_to_csv_categorical_and_ea(self):
+        # GH#46812
+        df = DataFrame({"a": "x", "b": [1, pd.NA]})
+        df["b"] = df["b"].astype("Int16")
+        df["b"] = df["b"].astype("category")
+        result = df.to_csv()
+        expected_rows = [",a,b", "0,x,1", "1,x,"]
+        expected = tm.convert_rows_list_to_csv_str(expected_rows)
+        assert result == expected
+
+    def test_to_csv_categorical_and_interval(self):
+        # GH#46297
+        df = DataFrame(
+            {
+                "a": [
+                    pd.Interval(
+                        Timestamp("2020-01-01"),
+                        Timestamp("2020-01-02"),
+                        closed="both",
+                    )
+                ]
+            }
+        )
+        df["a"] = df["a"].astype("category")
+        result = df.to_csv()
+        expected_rows = [",a", '0,"[2020-01-01, 2020-01-02]"']
+        expected = tm.convert_rows_list_to_csv_str(expected_rows)
+        assert result == expected