Python 正则表达式匹配多行字符串

Question

my_str:

PCT Filing Date: 2 December 2015
\nApplicants: Silixa Ltd.
\nChevron U.S.A. Inc. (Incorporated
in USA - California)
\nInventors: Farhadiroushan,
Mahmoud
\nGillies, Arran
Parker, Tom'

my code

regex = re.compile(r'(Applicants:)( )?(.*)', re.MULTILINE)
print(regex.findall(text))

my output:

[('Applicants:', ' ', 'Silixa Ltd.')]

what I need is to get the string between 'Applicants:' and '\nInventors:'

'Silixa Ltd.' & 'Chevron U.S.A. Inc. (Incorporated
in USA - California)'

Thanks in advance for your help

Answer 1

Try using re.DOTALL instead:

import re

text='''PCT Filing Date: 2 December 2015
\nApplicants: Silixa Ltd.
\nChevron U.S.A. Inc. (Incorporated
in USA - California)
\nInventors: Farhadiroushan,
Mahmoud
\nGillies, Arran
Parker, Tom'''

regex = re.compile(r'Applicants:(.*?)Inventors:', re.DOTALL)
print(regex.findall(text))

gives me

$ python test.py
[' Silixa Ltd.\n\nChevron U.S.A. Inc. (Incorporated\nin USA - California)\n\n']

The reason this works is that MULTILINE doesn't let the dot (.) match newlines, whereas DOTALL will.

Answer 2

If what you want is the contents between Applicants: and \nInventors: , your regex should reflect that:

>>> regex = re.compile(r'Applicants: (.*)Inventors:', re.S)
>>> print(regex.findall(s))
['Silixa Ltd.\n\nChevron U.S.A. Inc. (Incorporated\nin USA - California)\n']

re.S is the "dot matches all" option, so our (.*) will also match new lines. Note that this is different from re.MULTILINE , because re.MULTILINE only says that our expression should apply to multiple lines, but doesn't change the fact . will not match newlines. If . doesn't match newlines, a match like (.*) will still stop at newlines, not achieving the multiline effect you want.

Also note that if you are not interested in Applicants: or Inventors: you may not want to put that between () , as in (Inventors:) in your regex, because the match will try to create a matching group for it. That's the reason you got 3 elements in your output instead of just 1.

Answer 3

If you want to match all the text between \nApplicants: and \nInventors: , you could also get the match without using re.DOTALL preventing unnecessary backtracking.

Match Applicants: and capture in group 1 the rest of that same line and all lines that follow that do not start with Inventors:

Then match Inventors.

^Applicants: (.*(?:\r?\n(?!Inventors:).*)*)\r?\nInventors:

• ^ Start of string (Or use \b if it does not have to be at the start)
• Applicants: Match literally
• ( Capture group 1
  • .* Match the rest of the line
  • (?:\r?\n(?:Inventors.).*)* Match all lines that do not start with Inverntors:
• ) Close group
• \r?\nInventors: Match a newline and Inventors:

Regex demo | Python demo

Example code

import re
text = ("PCT Filing Date: 2 December 2015\n"
    "Applicants: Silixa Ltd.\n"
    "Chevron U.S.A. Inc. (Incorporated\n"
    "in USA - California)\n"
    "Inventors: Farhadiroushan,\n"
    "Mahmoud\n"
    "Gillies, Arran\n"
    "Parker, Tom'")
regex = re.compile(r'^Applicants: (.*(?:\r?\n(?!Inventors:).*)*)\r?\nInventors:', re.MULTILINE)
print(regex.findall(text))

Output

['Silixa Ltd.\nChevron U.S.A. Inc. (Incorporated\nin USA - California)']

Answer 4

Here is a more general approach to parse a string like that into a dict of all the keys and values in it (ie, any string at the start of a line followed by a : is a key and the string following that key is data):

import re 

txt="""\
PCT Filing Date: 2 December 2015
Applicants: Silixa Ltd.
Chevron U.S.A. Inc. (Incorporated
in USA - California)
Inventors: Farhadiroushan,
Mahmoud
Gillies, Arran
Parker, Tom'"""

pat=re.compile(r'(^[^\n:]+):[ \t]*([\s\S]*?(?=(?:^[^\n:]*:)|\Z))', flags=re.M)
data={m.group(1):m.group(2) for m in pat.finditer(txt)}

Result:

>>> data
{'PCT Filing Date': '2 December 2015\n', 'Applicants': 'Silixa Ltd.\nChevron U.S.A. Inc. (Incorporated\nin USA - California)\n', 'Inventors': "Farhadiroushan,\nMahmoud\nGillies, Arran\nParker, Tom'"}

>>> data['Applicants']
'Silixa Ltd.\nChevron U.S.A. Inc. (Incorporated\nin USA - California)\n'

Demo of the regex