Describe the issue:

np.quantile with weights being all zeros should probably give a error/warning. It's effectively like asking for a quantile of an empty array. Currently numpy seems to return the first sample.

Reproduce the code example:

np.quantile([1,2,3,4], 0.5, weights=[0,0,0,0], method='inverted_cdf')

Error message:

/.../numpy/lib/_function_base_impl.py:4858: RuntimeWarning: invalid value encountered in divide
  cdf /= cdf[-1, ...]  # normalization to 1
Out[9]: array(1)

Python and NumPy Versions:

2.1.3
3.13.0 | packaged by conda-forge | (main, Nov 27 2024, 19:18:50) [GCC 13.3.0]

Runtime Environment:

No response

Context for the issue:

Context: Noticed when looking into scikit-learn/scikit-learn#31032