BUG: Inconsistent conversion of object-type bytes array to StringDType #28269
Description
Describe the issue:
Converting an array of object dtype and bytes items, or a list of bytes, to variable-width strings results in a spurious b'...' tag.
This issue is absent in any other combination of input/output type where the input is bytes and the output is string.
Reproduce the code example:
>>> import numpy as np
>>> np.array(["foo", "bar"], dtype="O").astype("U")
array(['foo', 'bar'], dtype='<U3')
>>> np.array(["foo", "bar"], dtype="O").astype("T")
array(['foo', 'bar'], dtype=StringDType())
>>> np.array([b"foo", b"bar"], dtype="O").astype("U")
array(['foo', 'bar'], dtype='<U3')
>>> np.array([b"foo", b"bar"], dtype="S").astype("T")
array(['foo', 'bar'], dtype=StringDType())
>>> np.array([b"foo", b"bar"], dtype="T")
array(["b'foo'", "b'bar'"], dtype=StringDType())
>>> np.array([b"foo", b"bar"], dtype="O").astype("T")
array(["b'foo'", "b'bar'"], dtype=StringDType())Python and NumPy Versions:
numpy 2.2.2