Backport PR #55366 on branch 2.1.x (BUG: Inserting ndim=0 array does …

…not infer string dtype) (#55396)

Backport PR #55366: BUG: Inserting ndim=0 array does not infer string dtype

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

doc/source/whatsnew/v2.1.2.rst

@@ -22,6 +22,7 @@ Fixed regressions
 Bug fixes
 ~~~~~~~~~
 - Fixed bug in :meth:`Categorical.equals` if other has arrow backed string dtype (:issue:`55364`)
+- Fixed bug in :meth:`DataFrame.__setitem__` not inferring string dtype for zero-dimensional array with ``infer_string=True`` (:issue:`55366`)
 - Fixed bug in :meth:`DataFrame.idxmin` and :meth:`DataFrame.idxmax` raising for arrow dtypes (:issue:`55368`)
 - Fixed bug in :meth:`Index.insert` raising when inserting ``None`` into :class:`Index` with ``dtype="string[pyarrow_numpy]"`` (:issue:`55365`)
 - Silence ``Period[B]`` warnings introduced by :issue:`53446` during normal plotting activity (:issue:`55138`)

pandas/core/construction.py

@@ -562,7 +562,12 @@ def sanitize_array(
     if not is_list_like(data):
         if index is None:
             raise ValueError("index must be specified when data is not list-like")
+        if isinstance(data, str) and using_pyarrow_string_dtype():
+            from pandas.core.arrays.string_ import StringDtype
+
+            dtype = StringDtype("pyarrow_numpy")
         data = construct_1d_arraylike_from_scalar(data, len(index), dtype)
+
         return data
 
     elif isinstance(data, ABCExtensionArray):

pandas/tests/frame/indexing/test_indexing.py

@@ -1905,6 +1905,19 @@ def test_adding_new_conditional_column() -> None:
     tm.assert_frame_equal(df, expected)
 
 
+def test_add_new_column_infer_string():
+    # GH#55366
+    pytest.importorskip("pyarrow")
+    df = DataFrame({"x": [1]})
+    with pd.option_context("future.infer_string", True):
+        df.loc[df["x"] == 1, "y"] = "1"
+    expected = DataFrame(
+        {"x": [1], "y": Series(["1"], dtype="string[pyarrow_numpy]")},
+        columns=Index(["x", "y"], dtype="string[pyarrow_numpy]"),
+    )
+    tm.assert_frame_equal(df, expected)
+
+
 class TestSetitemValidation:
     # This is adapted from pandas/tests/arrays/masked/test_indexing.py
     # but checks for warnings instead of errors.