ENH: Enable assert_almost_equal to test numpy-like zero-dimensional arrays

[andrewgsavage]

Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

is_list_like was updated to return False for scalars of numpy-like array implementations. #35131 This has helped pass more extensionarray in pint-pandas .

Some tests now fail at assert_..._equal instead of at is_list_like, as isiterable in assert_almost_equal recognises pint-pandas' scalar, a Quantity, as iterable since it has an __iter__ method. The Quantity will TypeError since it isn't iterable.

cdef bint isiterable(obj):
    return hasattr(obj, '__iter__')

I've had a go at this #48920

Feature Description

Add a check for ndim == 0, like was done for is_list_like

cdef bint isiterable(obj):
    if hasattr(obj, '__iter__'):
        if hasattr(obj, "ndim") and obj.ndim == 0:
            return False
        return True
    return False

Alternative Solutions

Currently these tests are xpassed.

Additional Context

No response

[kvn4]

take

[kvn4]

Hey @andrewgsavage I am taking this issue. I was wondering why your original CR wasn't accepted. What needed to be changed?

[andrewgsavage]

I made requested changes and was waiting on review.

[kvn4]

@andrewgsavage Did you still want to make the changes and submit another CR? If not, do you mind if I make them for you?

[andrewgsavage]

please be my guest. I don't recall it waiting on any changes.

…

On Mon, Aug 7, 2023 at 9:32 PM Kevin ***@***.***> wrote: @andrewgsavage <https://github.com/andrewgsavage> Did you still want to make the changes and submit another CR? If not, do you mind if I make them for you? — Reply to this email directly, view it on GitHub <#48919 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADEMLEGRYLDV37WZZXUNLT3XUFGFVANCNFSM6AAAAAAQ3EKAZM> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[kvn4]

Hi @andrewgsavage, in your test:
(MockNumpyLikeArray(np.ndarray((2,) * 1)), "duck-ndarray-1d"), (MockNumpyLikeArray(np.array([])), "duck-ndarray-1d-empty"),
why aren't these considered iterable?

[andrewgsavage]

Try iterating them:

TypeError                                 Traceback (most recent call last)
----> 1 next(np.ndarray((2,) * 1))

TypeError: 'numpy.ndarray' object is not an iterator

TypeError                                 Traceback (most recent call last)
Cell In[5], line 1
----> 1 next(np.array([]))

TypeError: 'numpy.ndarray' object is not an iterator