median raises assertion error for an empty array

[adityav1810]

This is my first time contributing to an opensource project.
Solving issue #9433
It would be great if someone could help me figure out where can i add some test cases to support this solution.

[kc611]

Hi @adityav1810, Thank you for your contribution to Numba. 😄

I think the tests you're looking for is in numba/tests/test_array_reductions.py::TestArrayReductions::check_median_basic, the actual test itself (on line 343 of same file) is quite confusing, so you'd want to simply add a empty array case in this particular function and run the check function on it.

Also do remember to run flake8 on the file that you change.

Feel free to ask questions if any.

[adityav1810]

Hi @kc611 . Thanks for the help. I added the test and ran flake8 as well. However, it seems like the tests fail. Anyidea what I am missing?

[guilhermeleobas]

You cannot call np.median when the array dtype is datetime* and it is empty:

>>> import numpy as np
>>> a = np.array([]).astype(np.datetime64)
>>> np.median(a)
/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/core/fromnumeric.py:3474: RuntimeWarning: Mean of empty slice.
  return _methods._mean(a, axis=axis, dtype=dtype,
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "<__array_function__ internals>", line 180, in median
  File "/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/lib/function_base.py", line 3793, in median
    r, k = _ureduce(a, func=_median, axis=axis, out=out,
  File "/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/lib/function_base.py", line 3702, in _ureduce
    r = func(a, **kwargs)
  File "/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/lib/function_base.py", line 3847, in _median
    rout = mean(part[indexer], axis=axis, out=out)
  File "<__array_function__ internals>", line 180, in mean
  File "/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/core/fromnumeric.py", line 3474, in mean
    return _methods._mean(a, axis=axis, dtype=dtype,
  File "/Users/guilhermeleobas/micromamba/envs/numba/lib/python3.9/site-packages/numpy/core/_methods.py", line 179, in _mean
    ret = umr_sum(arr, axis, dtype, out, keepdims, where=where)
numpy.core._exceptions._UFuncBinaryResolutionError: ufunc 'add' cannot use operands with types dtype('<M8') and dtype('<M8')

Changing your code to the one below should fix the issue:

@overload(np.median)
def np_median(a):
    if not isinstance(a, types.Array):
        return

    is_datetime = as_dtype(a.dtype).char in 'mM'

    def median_impl(a):
        # np.median() works on the flattened array, and we need a temporary
        # workspace anyway
        temp_arry = a.flatten()
        n = temp_arry.shape[0]
        if not is_datetime and n == 0:
            return np.nan
        return _median_inner(temp_arry, n)
    return median_impl

[guilhermeleobas]

Regarding the datetime64 failure, I think this may be a bug on NumPy:
numpy/numpy#26216

[guilhermeleobas]

@adityav1810 Can you add a release notes file to your PR?
https://numba.readthedocs.io/en/latest/developer/contributing.html#release-notes