I tried to write python script to retrieve year and month from string.

The requirements are,

• year is fixed at 4 characters
• month is allowed to be two consecutive characters, or one or two characters when followed by a non-numeric value

"""
This is the "search year and month" module.

>>> search_year_month('202301')
True
>>> search_year_month('2023-1')
True
>>> search_year_month('2023-01')
True
>>> search_year_month('20231')
False
"""

import re

_re = re.compile(
    r"(?P<year>\d{4})"
    r"(?P<month>\d{2}|(?<=[^\d])\d{1,2})"
)

def search_year_month(v):
    match = _re.search(v)
    return match is not None

if __name__ == "__main__":
    import doctest
    doctest.testmod()

But, 2023-1 and 2023-01 are failed... Is there a better way to build regular expressions?

I tried only month part, I got expected result.

"""
This is the "single lookbehind sample" module.

>>> search_lookbehind('01')
True
>>> search_lookbehind('-1')
True
>>> search_lookbehind('-01')
True
>>> search_lookbehind('1')
False
"""

import re

_re = re.compile(
    r"(?P<month>\d{2}|(?<=[^\d])\d{1,2})"
)

def search_lookbehind(v):
    match = _re.search(v)
    return match is not None

if __name__ == "__main__":
    import doctest
    doctest.testmod()

Solution

I separated regex by each section, and applied search and replace strategy.

... Why does regular expression splitting work and concatenation not?

"""
This is the "search year and month" module.

>>> search_year_month('202301')
True
>>> search_year_month('2023-1')
True
>>> search_year_month('2023-01')
True
>>> search_year_month('20231')
False

>>> retrieve_year_month('202301')
(2023, 1)
>>> retrieve_year_month('2023-1')
(2023, 1)
>>> retrieve_year_month('2023-01')
(2023, 1)
>>> retrieve_year_month('20231')

"""

import re

fields = ["year", "month"]

_re = {
    "year": re.compile(r"^(?P<year>\d{4})"),
    "month": re.compile(r"(?P<month>(?:\d{2}|(?<=[^\d])\d{1,2}))"),
}


def retrieve_matched_value(match, s):
    head, tail = match.span()
    return s[head:tail], s[tail:]


def search_year_month(s):
    for field in fields:
        match = _re[field].search(s)
        if match is None:
            return False
        _, s = retrieve_matched_value(match, s)
    return True


def retrieve_year_month(s):
    result = tuple()
    for field in fields:
        match = _re[field].search(s)
        if match is None:
            return None
        value, s = retrieve_matched_value(match, s)
        result = result   (int(value),)
    return result


if __name__ == "__main__":
    import doctest

    doctest.testmod()

CodePudding user response：

option 1: search and replace strategy

import re
test_string=['202301','2023-1','2023-01','20231']
year=r"^\d{4}"
sep=r"\W "
for e in test_string:
    year=re.findall(r"^\d{4}",'202301')[0]
    e=re.sub(year,"",e)
    e=re.sub(sep,"",e)
    print(f"year:{year},month:{e}")

option 2: regex (all in one)

re_all=r"(^\d{4})(?:\W ){0,1}(\d )"
for e in test_string:
    search_results=re.search(re_all,e)
    print(f"year:{search_results.group(1)},month:{search_results.group(2)}")

same outcome:

year:2023,month:01
year:2023,month:1
year:2023,month:01
year:2023,month:1

CodePudding user response：

You could use a conditional checking for group 2 that holds an optional - char.

If there is a value for group 2, then match 1 or 2 digits for the month, else match 2 digits for the month.

\b(?P<year>\d{4})(-)?(?P<month>(?(2)\d\d?|\d\d))\b

Explanation

• \b A word boundary to prevent a partial word match
• (?P<year>\d{4}) Group year matching 4 digits
• (-)? Optional capture group 2 matching a -
• (?P<month> Group month
  • (? Start a conditional
    • (2)\d\d? If we have group 2, match 1 or 2 digits
    • | Or
    • \d\d Match 2 digits
  • ) Close the conditional
• ) Close group month
• \b A word boundary

See a regex101 demo and a Python demo.

For example

import re

pattern = r"\b(?P<year>\d{4})(-)?(?P<month>(?(2)\d\d?|\d\d))\b"
s = ("202301\n"
            "2023-1\n"
            "2023-01\n"
            "20231")

for m in re.finditer(pattern, s):
    print(m.groupdict())

Output

{'year': '2023', 'month': '01'}
{'year': '2023', 'month': '1'}
{'year': '2023', 'month': '01'}