BUG: np.where may not preserve Enum type when used with a mix of Enums defined using a custom class that inherits from both str and Enum

[luke396]

Describe the issue:

Hi, when using the np.where function with an array of Enums defined using a custom class that inherits from both str and Enum, the resulting array does not preserve the Enum type as expected.

Reproduce the code example:

import numpy as np
from enum import Enum, auto



class TestEnum(str, Enum):
    first = auto()
    second = auto()


cond = np.array([True, False, False])
a = np.array([TestEnum.first, None, None])
b = TestEnum.second
np.where(cond, a, b) 
# now: array([<TestEnum.first: '1'>, 'T', 'T'], dtype=object)
# expected: array([<TestEnum.first: 1>, <TestEnum.second: 2>, <TestEnum.second: 2>], dtype=object)

When class TestEnum(Enum):, it works well.

Error message:

No response

Runtime information:

1.24.0
3.8.15 | packaged by conda-forge | (default, Nov 22 2022, 08:49:35)
[GCC 10.4.0]

Context for the issue:

xref BUG: enum not properly handled by fillna · Issue #51941 · pandas-dev/pandas

[maniben3]

import numpy as np
from enum import Enum, auto

class TestEnum(str, Enum):
first = auto()
second = auto()

cond = np.array([True, False, False])
a = np.array([TestEnum.first, None, None])
b = (TestEnum.second)
b=np.array(b,dtype=np.object)
np.where(cond, a, b)

:14: DeprecationWarning: np.object is a deprecated alias for the builtin object. To silence this warning, use object by itself. Doing this will not modify any behavior and is safe.
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
b=np.array(b,dtype=np.object)
array([<TestEnum.first: '1'>, <TestEnum.second: '2'>,
<TestEnum.second: '2'>], dtype=object)

#use with np.object

[luke396]

b = np.array(TestEnum.second) indeed works.