seems that from 1.23 the 'delimiter' option of loadtxt require single unicode character #22311
Description
Describe the issue:
It seems that starting from 1.23 the 'delimiter' option of loadtxt requires a single unicode character where a random string was allowed before (and the doc of 1.23 doesn't tell about this new behavior)
So, when running code that worked fine up before 1.23, we get the error whose stack trace is below.
Reproduce the code example:
import numpy as np
np.loadtxt(pos_file,comments="@",delimiter="val=",usecols=(0,))
where the text file is something like:
------
field1val=12val=13val=14
field2val=22val=23val=24
------Error message:
Traceback (most recent call last):
File "", line 114, in <module>
pos_data = np.loadtxt(pos_file,comments="@",delimiter="val=",usecols=(0,))
File "/opt/local/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/site-packages/numpy/lib/npyio.py", line 1313, in loadtxt
arr = _read(fname, dtype=dtype, comment=comment, delimiter=delimiter,
File "/opt/local/Library/Frameworks/Python.framework/Versions/3.8/lib/python3.8/site-packages/numpy/lib/npyio.py", line 979, in _read
arr = _load_from_filelike(
TypeError: Text reading control character must be a single unicode character or None; but got: 'val='NumPy/Python version information:
1.23.0
Context for the issue:
j