Regex with lookahead does not match in Python

Question

I have composed a regex pattern aiming to capture one date and one number from a sentence. But it does not.

My code is:

txt = 'Την 02/12/2013 καταχωρήθηκε στο Γενικό Εμπορικό Μητρώο της Υπηρεσίας Γ.Ε.ΜΗ. του Επιμελητηρίου Βοιωτίας, με κωδικόαριθμό καταχώρισης Κ.Α.Κ.: 110035'

p = re.compile(r'''Την\s? # matches Την with a possible space afterwards

               (?P\d{2}/\d{2}/\d{4}) #matches a date of the given format and captures it with a named group
               
               \.+ # Allow for an arbitrary sequence of characters 
               
               (?=(κωδικ.\s?αριθμ.\s?καταχ.ριση.)|(κ\.?α\.?κ\.?:?\s*)) # defines two lookaheads, either of which suffices
               
               (?P\d+) # captures a sequence of numbers''', re.I|re.VERBOSE)

p.findall(txt)

I would expect to return a list with two elements: '02/12/2013' and '110035', but instead, it returns an empty list.

user_3832970 · Answer

Issues:

\.+ matches one or more dots, you need to use .+ (no escaping)
(?=(κωδικ.\s?αριθμ.\s?καταχ.ριση.)|(κ\.?α\.?κ\.?:?\s*))(?P\d+) will always prevent any match since the positive lookahead requires some text that is not 1 or more digits. You need to convert the lookahead to a consuming pattern.

I suggest fixing your pattern as

p = re.compile(r'''Την\s? # matches Την with a possible space afterwards
(?P\d{2}/\d{2}/\d{4}) #matches a date of the given format and captures it with a named group
.+ # Allow for an arbitrary sequence of characters 
(?:κωδικ.\s?αριθμ.\s?καταχ.ριση.|κ\.?α\.κ\.:?)\s+ # defines two lookaheads, either of which suffices
(?P\d+) # captures a sequence of numbers''', re.I | re.X)

See the regex demo

Details

Την\s? - Την string and an optional whitespace
(?P\d{2}/\d{2}/\d{4}) - Group "KEK_date": a date pattern, 2 digits, /, 2 digits, / and 4 digits
.+ - 1 or more chars other than line break chars as many as possible
(?:κωδικ.\s?αριθμ.\s?καταχ.ριση.|κ\.?α\.κ\.:?) - either of
- κωδικ.\s?αριθμ.\s?καταχ.ριση. - κωδικ, any char, an optional whitespace, αριθμ, any one char, an optional whitespace, καταχ, any 1 char, ριση and any 1 char (but line break char)
- | - or
- κ\.?α\.κ\.:? - κ, an optional ., α, an optional ., κ a . and then an optional :
\s+ - 1+ whitespaces
(?P\d+) - Group "KEK_number": 1+ digits

See a Python demo:

import re
txt = 'Την 02/12/2013 καταχωρήθηκε στο Γενικό Εμπορικό Μητρώο της Υπηρεσίας Γ.Ε.ΜΗ. του Επιμελητηρίου Βοιωτίας, με κωδικόαριθμό καταχώρισης Κ.Α.Κ.: 110035'
p = re.compile(r'''Την\s? # matches Την with a possible space afterwards
(?P\d{2}/\d{2}/\d{4}) #matches a date of the given format and captures it with a named group
.+ # Allow for an arbitrary sequence of characters 
(?:κωδικ.\s?αριθμ.\s?καταχ.ριση.|κ\.?α\.κ\.:?)\s+ # defines two lookaheads, either of which suffices
(?P\d+) # captures a sequence of numbers''', re.I | re.X)
print(p.findall(txt)) # => [('02/12/2013', '110035')]

Regex with lookahead does not match in Python

1 Answers