updated cast_scalar_to_arr to support tuple shape for extension dtype

Author: justinessert

pandas/core/dtypes/cast.py

@@ -758,6 +758,10 @@ def infer_dtype_from_array(arr, pandas_dtype: bool = False) -> Tuple[DtypeObj, A
     if pandas_dtype and is_extension_array_dtype(arr):
         return arr.dtype, arr
 
+    dtype, _ = infer_dtype_from_scalar(arr[0], pandas_dtype=True)
+    if is_extension_array_dtype(dtype):
+        return dtype, arr
+
     elif isinstance(arr, ABCSeries):
         return arr.dtype, np.asarray(arr)
 
@@ -1510,7 +1514,10 @@ def cast_scalar_to_array(shape, value, dtype: Optional[DtypeObj] = None) -> np.n
         fill_value = value
 
     if is_extension_array_dtype(dtype):
-        values = dtype.construct_array_type()._from_sequence([value] * shape)
+        if isinstance(shape, int):
+            shape = (shape, 1)
+        value = [construct_1d_arraylike_from_scalar(value, shape[0], dtype)]
+        values = value * shape[1]
     else:
         values = np.empty(shape, dtype=dtype)
         values.fill(fill_value)

pandas/core/frame.py

@@ -76,6 +76,7 @@
     cast_scalar_to_array,
     coerce_to_dtypes,
     find_common_type,
+    infer_dtype_from_array,
     infer_dtype_from_scalar,
     invalidate_string_dtypes,
     maybe_cast_to_datetime,
@@ -528,9 +529,15 @@ def __init__(
                 values = cast_scalar_to_array(
                     (len(index), len(columns)), data, dtype=dtype
                 )
-                mgr = init_ndarray(
-                    values, index, columns, dtype=values.dtype, copy=False
-                )
+                if isinstance(values, list):
+                    # Case 1: values is a list of extension arrays
+                    dtype, _ = infer_dtype_from_array(values[0], pandas_dtype=True)
+                    mgr = arrays_to_mgr(values, columns, index, columns, dtype=dtype)
+                else:
+                    # Case 2: values is a numpy array
+                    mgr = init_ndarray(
+                        values, index, columns, dtype=values.dtype, copy=False
+                    )
             else:
                 raise ValueError("DataFrame constructor not properly called!")
 
@@ -3731,6 +3738,11 @@ def reindexer(value):
 
             # upcast
             value = cast_scalar_to_array(len(self.index), value)
+
+            # if extension dtype, value will be a list of length 1
+            if isinstance(value, list):
+                value = value[0]
+
             value = maybe_cast_to_datetime(value, infer_dtype)
 
         # return internal types directly

pandas/tests/dtypes/cast/test_infer_dtype.py

@@ -214,4 +214,4 @@ def test_cast_scalar_to_extension_array(obj, dtype):
     exp = dtype.construct_array_type()._from_sequence([obj] * shape)
 
     arr = cast_scalar_to_array(shape, obj, dtype=dtype)
-    tm.assert_extension_array_equal(arr, exp)
+    tm.assert_extension_array_equal(arr[0], exp)