csvのデリミタ、改行文字とかを推測

csv定義　推測関数

def sniffer(dat:str,
    delimiters = ['\t', ';', ',', ':', '|', ' '],
    quotechars = ['"', '\'', '`'],
    escapechars = ['\\', '^'],
    ):

    # guess lineterminator
    LF = ""
    lines = 1
    if ('\r' in dat):
        LF += '\r'
    if ('\n' in dat):
        LF += '\n'
        lines = dat.count('\n')
    if (LF == "" and '0x0' in dat):
        LF = '0x0'
        lines = dat.count('0x0')
    
    # guess delimiter
    mxdlm = 0
    delimiter = None
    _it_delim = ((dat.count(x), x) for x in delimiters if x in dat)
    try: mxdlm, delimiter = max((v, x) for v, x in _it_delim if (v / lines) >= 1)
    except ValueError: pass

    # guess quotes
    mxqt = 0
    quotechar = '"'
    _it_quot = ((dat.count(x), x) for x in quotechars if x in dat)
    try: mxqt, quotechar = max((v, x) for v, x in _it_quot if v % 2 == 0)
    except ValueError: pass

    # guess escapechar, etc
    if quotechar:
        escapechar, doublequote = next(((esc, False) for esc in escapechars if (esc + quotechar) in dat), (None, True))

    return dict(
        delimiter=delimiter,
        doublequote=doublequote,
        lineterminator=LF,
        quoting = 1 if quotechar and mxqt > mxdlm * 2 else 0,
        quotechar=quotechar,
        skipinitialspace = delimiter and (delimiter + ' ') in dat)

使い方

>>> csvdat = """n,aa
1,1
2,あ
"""
>>> print(sniffer(csvdat))
{'delimiter': ',', 'doublequote': True, 'escapechar': None, 'lineterminator': '\n', 'quotechar': '"', 'quoting': 0, 'skipinitialspace': False}

>>> import csv
>>> from io import StringIO
>>> r = csv.reader(StringIO(csvdat), **sniffer(csvdat))
>>> print(list(r))
[['n', 'aa'], ['1', '1'], ['2', 'あ']]

以上csv.Sniffer 車輪の再発明