REF: simplify try_cast (pandas-dev#41811)

Author: jbrockmendel

pandas/core/construction.py

@@ -32,7 +32,6 @@
 )
 from pandas.core.dtypes.cast import (
     construct_1d_arraylike_from_scalar,
-    construct_1d_ndarray_preserving_na,
     construct_1d_object_array_from_listlike,
     maybe_cast_to_datetime,
     maybe_cast_to_integer_array,
@@ -726,6 +725,10 @@ def _try_cast(
             return subarr
         return ensure_wrapped_if_datetimelike(arr).astype(dtype, copy=copy)
 
+    elif dtype.kind == "U":
+        # TODO: test cases with arr.dtype.kind in ["m", "M"]
+        return lib.ensure_string_array(arr, convert_na_value=False, copy=copy)
+
     elif dtype.kind in ["m", "M"]:
         return maybe_cast_to_datetime(arr, dtype)
 
@@ -735,16 +738,12 @@ def _try_cast(
         if is_integer_dtype(dtype):
             # this will raise if we have e.g. floats
 
-            maybe_cast_to_integer_array(arr, dtype)
-            subarr = arr
+            subarr = maybe_cast_to_integer_array(arr, dtype)
         else:
-            subarr = arr
-
-        if not isinstance(subarr, ABCExtensionArray):
             # 4 tests fail if we move this to a try/except/else; see
             #  test_constructor_compound_dtypes, test_constructor_cast_failure
             #  test_constructor_dict_cast2, test_loc_setitem_dtype
-            subarr = construct_1d_ndarray_preserving_na(subarr, dtype, copy=copy)
+            subarr = np.array(arr, dtype=dtype, copy=copy)
 
     except (ValueError, TypeError):
         if raise_cast_failure:

pandas/core/dtypes/cast.py

@@ -14,7 +14,6 @@
 from typing import (
     TYPE_CHECKING,
     Any,
-    Sequence,
     Sized,
     cast,
     overload,
@@ -1971,59 +1970,6 @@ def construct_1d_object_array_from_listlike(values: Sized) -> np.ndarray:
     return result
 
 
-def construct_1d_ndarray_preserving_na(
-    values: Sequence, dtype: np.dtype | None = None, copy: bool = False
-) -> np.ndarray:
-    """
-    Construct a new ndarray, coercing `values` to `dtype`, preserving NA.
-
-    Parameters
-    ----------
-    values : Sequence
-    dtype : numpy.dtype, optional
-    copy : bool, default False
-        Note that copies may still be made with ``copy=False`` if casting
-        is required.
-
-    Returns
-    -------
-    arr : ndarray[dtype]
-
-    Examples
-    --------
-    >>> np.array([1.0, 2.0, None], dtype='str')
-    array(['1.0', '2.0', 'None'], dtype='<U4')
-
-    >>> construct_1d_ndarray_preserving_na([1.0, 2.0, None], dtype=np.dtype('str'))
-    array(['1.0', '2.0', None], dtype=object)
-    """
-
-    if dtype is not None and dtype.kind == "U":
-        subarr = lib.ensure_string_array(values, convert_na_value=False, copy=copy)
-    else:
-        if dtype is not None:
-            _disallow_mismatched_datetimelike(values, dtype)
-
-        if (
-            dtype == object
-            and isinstance(values, np.ndarray)
-            and values.dtype.kind in ["m", "M"]
-        ):
-            # TODO(numpy#12550): special-case can be removed
-            subarr = construct_1d_object_array_from_listlike(list(values))
-        elif (
-            dtype is not None
-            and dtype.kind in ["i", "u"]
-            and isinstance(values, np.ndarray)
-            and values.dtype.kind == "f"
-        ):
-            return astype_float_to_int_nansafe(values, dtype, copy=copy)
-        else:
-            subarr = np.array(values, dtype=dtype, copy=copy)
-
-    return subarr
-
-
 def maybe_cast_to_integer_array(
     arr: list | np.ndarray, dtype: np.dtype, copy: bool = False
 ) -> np.ndarray:

pandas/tests/dtypes/cast/test_construct_ndarray.py

@@ -1,23 +1,22 @@
 import numpy as np
 import pytest
 
-from pandas.core.dtypes.cast import construct_1d_ndarray_preserving_na
-
 import pandas._testing as tm
+from pandas.core.construction import sanitize_array
 
 
 @pytest.mark.parametrize(
     "values, dtype, expected",
     [
-        ([1, 2, 3], None, np.array([1, 2, 3])),
+        ([1, 2, 3], None, np.array([1, 2, 3], dtype=np.int64)),
         (np.array([1, 2, 3]), None, np.array([1, 2, 3])),
         (["1", "2", None], None, np.array(["1", "2", None])),
         (["1", "2", None], np.dtype("str"), np.array(["1", "2", None])),
         ([1, 2, None], np.dtype("str"), np.array(["1", "2", None])),
     ],
 )
 def test_construct_1d_ndarray_preserving_na(values, dtype, expected):
-    result = construct_1d_ndarray_preserving_na(values, dtype=dtype)
+    result = sanitize_array(values, index=None, dtype=dtype)
     tm.assert_numpy_array_equal(result, expected)
 
 
@@ -27,5 +26,5 @@ def test_construct_1d_ndarray_preserving_na_datetimelike(dtype):
     expected = np.array(list(arr), dtype=object)
     assert all(isinstance(x, type(arr[0])) for x in expected)
 
-    result = construct_1d_ndarray_preserving_na(arr, np.dtype(object))
+    result = sanitize_array(arr, index=None, dtype=np.dtype(object))
     tm.assert_numpy_array_equal(result, expected)