Describe the issue:

numpy.result_dtype currently returns a dtype with length information when combining concrete fixed-width dtypes with their variable-length python equivalent:

np.result_type(np.dtype("<U4"), str)  # np.dtype("<U4")
np.result_type(np.dtype("S3"), bytes)  # np.dtype("S3")

I would argue that in this specific case result_type should instead drop the length information (i.e. return np.dtype("<U") and np.dtype("S")), since the python string may be longer than the current length of the dtype.

Obviously, for new code the new string dtype will not have this issue and should be preferred, but this concerns code that tries to be compatible with both (in my case, xarray).

cc @shoyer

Reproduce the code example:

import numpy as np

result = np.result_type(np.dtype("<U3"), str)
assert result == np.dtype("<U")  # assertion error

Error message:

No response

Python and NumPy Versions:

2.1.1
3.12.6 | packaged by conda-forge | (main, Sep 22 2024, 14:16:49) [GCC 13.3.0]

Runtime Environment:

No response

Context for the issue:

No response