Bug report

csv.Sniffer().sniff() detects the wrong field delimiter if the possible valid delimiters contain \t and the provided data contains one line starting with the combination of double quotes followed by a tab and has the same combination of double quotes and a tab at least once more afterwards in the data. In that case the delimiter is always reported as \t instead of the correct one.

Configuring the preferred separator as suggested in #80678 (comment) doesn't have any effect on this behavior.

Here is an example to reproduce it:

import csv
from pprint import pprint

data = """
Surname;First Name;Year of birth
"\tDoe;Jane"\t;1971
"""

sniffer = csv.Sniffer()
dialect = sniffer.sniff(data, delimiters=",;\t")
pprint(dict(dialect.__dict__))

Expected output:

{'__doc__': None,
 '__module__': 'csv',
 '_name': 'sniffed',
 'delimiter': ';',
 'doublequote': False,
 'lineterminator': '\r\n',
 'quotechar': '"',
 'quoting': 0,
 'skipinitialspace': False}

Actual output:

{'__doc__': None,
 '__module__': 'csv',
 '_name': 'sniffed',
 'delimiter': '\t',
 'doublequote': False,
 'lineterminator': '\r\n',
 'quotechar': '"',
 'quoting': 0,
 'skipinitialspace': False}

Your environment

• CPython versions tested on: 3.7.14, 3.8.14, 3.9.14, 3.10.7
• Operating system and architecture: Debian/sid on amd64