update constructor copy test + do not copy in maybe_convert_objects?

Author: jorisvandenbossche

pandas/_libs/lib.pyx

@@ -762,6 +762,7 @@ cpdef ndarray[object] ensure_string_array(
             out = arr.astype(str).astype(object)
             out[arr.isna()] = na_value
             return out
+        arr = arr.to_numpy(dtype=object)
     elif not util.is_array(arr):
         arr = np.array(arr, dtype="object")
 
@@ -2735,17 +2736,13 @@ def maybe_convert_objects(ndarray[object] objects,
             from pandas.core.arrays.string_ import StringDtype
 
             dtype = StringDtype()
-            return dtype.construct_array_type()._from_sequence(
-                objects, dtype=dtype, copy=True
-            )
+            return dtype.construct_array_type()._from_sequence(objects, dtype=dtype)
 
         elif using_string_dtype() and is_string_array(objects, skipna=True):
             from pandas.core.arrays.string_ import StringDtype
 
             dtype = StringDtype(na_value=np.nan)
-            return dtype.construct_array_type()._from_sequence(
-                objects, dtype=dtype, copy=True
-            )
+            return dtype.construct_array_type()._from_sequence(objects, dtype=dtype)
 
         seen.object_ = True
     elif seen.interval_:

pandas/tests/frame/test_constructors.py

@@ -309,19 +309,28 @@ def test_1d_object_array_does_not_copy(self, using_infer_string):
                 # no numpy arrays to compare
                 pass
             else:
-                # TODO(infer_string): this should be fixed to still share memory?
-                assert not np.shares_memory(df[0].to_numpy(), arr)
+                assert np.shares_memory(df[0].to_numpy(), arr)
         else:
             assert np.shares_memory(df.values, arr)
 
         df = DataFrame(arr, dtype=object, copy=False)
         assert np.shares_memory(df.values, arr)
 
-    @pytest.mark.xfail(using_string_dtype(), reason="conversion copies")
-    def test_2d_object_array_does_not_copy(self):
+    def test_2d_object_array_does_not_copy(self, using_infer_string):
         # https://github.com/pandas-dev/pandas/issues/39272
         arr = np.array([["a", "b"], ["c", "d"]], dtype="object")
         df = DataFrame(arr, copy=False)
+        if using_infer_string:
+            if df[0].dtype.storage == "pyarrow":
+                # object dtype strings are converted to arrow memory,
+                # no numpy arrays to compare
+                pass
+            else:
+                assert np.shares_memory(df[0].to_numpy(), arr)
+        else:
+            assert np.shares_memory(df.values, arr)
+
+        df = DataFrame(arr, dtype=object, copy=False)
         assert np.shares_memory(df.values, arr)
 
     def test_constructor_dtype_list_data(self):