BUG: PandasArray._from_sequence with list-of-tuples (#39828)

Author: jbrockmendel

pandas/core/arrays/numpy_.py

@@ -18,6 +18,7 @@
 )
 from pandas.compat.numpy import function as nv
 
+from pandas.core.dtypes.cast import construct_1d_object_array_from_listlike
 from pandas.core.dtypes.dtypes import PandasDtype
 from pandas.core.dtypes.missing import isna
 
@@ -97,6 +98,14 @@ def _from_sequence(
             dtype = dtype._dtype
 
         result = np.asarray(scalars, dtype=dtype)
+        if (
+            result.ndim > 1
+            and not hasattr(scalars, "dtype")
+            and (dtype is None or dtype == object)
+        ):
+            # e.g. list-of-tuples
+            result = construct_1d_object_array_from_listlike(scalars)
+
         if copy and result is scalars:
             result = result.copy()
         return cls(result)

pandas/core/internals/blocks.py

@@ -207,10 +207,17 @@ def _check_ndim(self, values, ndim):
         if ndim is None:
             ndim = values.ndim
 
-        if self._validate_ndim and values.ndim != ndim:
+        if self._validate_ndim:
+            if values.ndim != ndim:
+                raise ValueError(
+                    "Wrong number of dimensions. "
+                    f"values.ndim != ndim [{values.ndim} != {ndim}]"
+                )
+        elif values.ndim > ndim:
+            # ExtensionBlock
             raise ValueError(
                 "Wrong number of dimensions. "
-                f"values.ndim != ndim [{values.ndim} != {ndim}]"
+                f"values.ndim > ndim [{values.ndim} > {ndim}]"
             )
         return ndim
 
@@ -2178,28 +2185,6 @@ def fillna(
             value, limit=limit, inplace=inplace, downcast=downcast
         )
 
-    def _check_ndim(self, values, ndim):
-        """
-        ndim inference and validation.
-
-        This is overridden by the DatetimeTZBlock to check the case of 2D
-        data (values.ndim == 2), which should only be allowed if ndim is
-        also 2.
-        The case of 1D array is still allowed with both ndim of 1 or 2, as
-        if the case for other EAs. Therefore, we are only checking
-        `values.ndim > ndim` instead of `values.ndim != ndim` as for
-        consolidated blocks.
-        """
-        if ndim is None:
-            ndim = values.ndim
-
-        if values.ndim > ndim:
-            raise ValueError(
-                "Wrong number of dimensions. "
-                f"values.ndim != ndim [{values.ndim} != {ndim}]"
-            )
-        return ndim
-
 
 class TimeDeltaBlock(DatetimeLikeBlockMixin):
     __slots__ = ()

pandas/tests/extension/test_numpy.py

@@ -189,11 +189,6 @@ def test_getitem_scalar(self, data):
         # AssertionError
         super().test_getitem_scalar(data)
 
-    @skip_nested
-    def test_take_series(self, data):
-        # ValueError: PandasArray must be 1-dimensional.
-        super().test_take_series(data)
-
 
 class TestGroupby(BaseNumPyTests, base.BaseGroupbyTests):
     def test_groupby_extension_apply(
@@ -222,13 +217,6 @@ def test_shift_fill_value(self, data):
         # np.array shape inference. Shift implementation fails.
         super().test_shift_fill_value(data)
 
-    @skip_nested
-    @pytest.mark.parametrize("box", [pd.Series, lambda x: x])
-    @pytest.mark.parametrize("method", [lambda x: x.unique(), pd.unique])
-    def test_unique(self, data, box, method):
-        # Fails creating expected
-        super().test_unique(data, box, method)
-
     @skip_nested
     def test_fillna_copy_frame(self, data_missing):
         # The "scalar" for this array isn't a scalar.
@@ -244,31 +232,10 @@ def test_searchsorted(self, data_for_sorting, as_series):
         # Test setup fails.
         super().test_searchsorted(data_for_sorting, as_series)
 
-    @skip_nested
-    def test_where_series(self, data, na_value, as_frame):
-        # Test setup fails.
-        super().test_where_series(data, na_value, as_frame)
-
-    @pytest.mark.parametrize("repeats", [0, 1, 2, [1, 2, 3]])
-    def test_repeat(self, data, repeats, as_series, use_numpy, request):
-        if data.dtype.numpy_dtype == object and repeats != 0:
-            mark = pytest.mark.xfail(reason="mask shapes mismatch")
-            request.node.add_marker(mark)
-        super().test_repeat(data, repeats, as_series, use_numpy)
-
     @pytest.mark.xfail(reason="PandasArray.diff may fail on dtype")
     def test_diff(self, data, periods):
         return super().test_diff(data, periods)
 
-    @pytest.mark.parametrize("box", [pd.array, pd.Series, pd.DataFrame])
-    def test_equals(self, data, na_value, as_series, box, request):
-        # Fails creating with _from_sequence
-        if box is pd.DataFrame and data.dtype.numpy_dtype == object:
-            mark = pytest.mark.xfail(reason="AssertionError in _get_same_shape_values")
-            request.node.add_marker(mark)
-
-        super().test_equals(data, na_value, as_series, box)
-
 
 class TestArithmetics(BaseNumPyTests, base.BaseArithmeticOpsTests):
     divmod_exc = None
@@ -289,8 +256,11 @@ def test_divmod_series_array(self, data):
     def test_arith_series_with_scalar(self, data, all_arithmetic_operators):
         super().test_arith_series_with_scalar(data, all_arithmetic_operators)
 
-    @skip_nested
-    def test_arith_series_with_array(self, data, all_arithmetic_operators):
+    def test_arith_series_with_array(self, data, all_arithmetic_operators, request):
+        opname = all_arithmetic_operators
+        if data.dtype.numpy_dtype == object and opname not in ["__add__", "__radd__"]:
+            mark = pytest.mark.xfail(reason="Fails for object dtype")
+            request.node.add_marker(mark)
         super().test_arith_series_with_array(data, all_arithmetic_operators)
 
     @skip_nested
@@ -325,11 +295,6 @@ def test_fillna_scalar(self, data_missing):
         # Non-scalar "scalar" values.
         super().test_fillna_scalar(data_missing)
 
-    @skip_nested
-    def test_fillna_series_method(self, data_missing, fillna_method):
-        # Non-scalar "scalar" values.
-        super().test_fillna_series_method(data_missing, fillna_method)
-
     @skip_nested
     def test_fillna_series(self, data_missing):
         # Non-scalar "scalar" values.
@@ -358,20 +323,6 @@ def test_merge(self, data, na_value):
         # Fails creating expected (key column becomes a PandasDtype because)
         super().test_merge(data, na_value)
 
-    @skip_nested
-    def test_merge_on_extension_array(self, data):
-        # Fails creating expected
-        super().test_merge_on_extension_array(data)
-
-    @skip_nested
-    def test_merge_on_extension_array_duplicates(self, data):
-        # Fails creating expected
-        super().test_merge_on_extension_array_duplicates(data)
-
-    @skip_nested
-    def test_transpose_frame(self, data):
-        super().test_transpose_frame(data)
-
 
 class TestSetitem(BaseNumPyTests, base.BaseSetitemTests):
     @skip_nested