7.23. Regex Backreference

\g<number> - backreferencing by position
\g<name> - backreferencing by name
(?P=name) - backreferencing by name
Note, that for backreference, must use raw-sting or double backslash

7.23.1. SetUp

>>> import re

7.23.2. Backreference

>>> string = '<p>Hello World</p>'
>>> pattern = r'<(?P<tag>.+)>(?:.+)</(?P=tag)>'
>>>
>>> re.findall(pattern, string)
['p']

7.23.3. Recall Group by Position

\g<number> - backreferencing by position

>>> string = '<p>Hello World</p>'
>>> pattern = r'<p>(.+)</p>'
>>> replace = r'<strong>\g<1></strong>'
>>>
>>> re.sub(pattern, replace, string)
'<strong>Hello World</strong>'

7.23.4. Recall Group by Name

\g<name> - backreferencing by name

>>> string = '<p>Hello World</p>'
>>> pattern = r'<p>(?P<text>.+)</p>'
>>> replace = r'<strong>\g<text></strong>'
>>>
>>> re.sub(pattern, replace, string)
'<strong>Hello World</strong>'

7.23.5. Example

(?P<tag><.*?>).+(?P=tag) - matches string inside of a <tag> (opening and closing tag is the same)

7.23.6. Use Case - 1

>>> import re
>>>
>>> string = 'On Sun, Jan 1st, 2000 at 12:00 AM Alice <alice@example.com> wrote'
>>>
>>> year = r'(?P<year>\d{4})'
>>> month = r'(?P<month>[A-Z][a-z]{2})'
>>> day = r'(?P<day>\d{1,2})'
>>> pattern = f'{month} {day}(?:st|nd|rd|th), {year}'

Recall group by position:

>>> replace = r'\g<3> \g<1> \g<2>'
>>>
>>> re.sub(pattern, replace, string)
'On Sun, 2000 Jan 1 at 12:00 AM Alice <alice@example.com> wrote'

Recall group by name:

>>> replace = r'\g<year> \g<month> \g<day>'
>>>
>>> re.sub(pattern, replace, string)
'On Sun, 2000 Jan 1 at 12:00 AM Alice <alice@example.com> wrote'

7.23.7. Use Case - 2

>>> import re
>>>
>>>
>>> string = '<p>Hello World</p>'
>>> pattern = r'<(?P<tag>.*?)>(.*?)</(?P=tag)>'
>>>
>>> re.findall(pattern, string)
[('p', 'Hello World')]

7.23.8. Use Case - 3

>>> import re
>>>
>>>
>>> string = '<p>We choose to go to the <strong>Moon</strong></p>'
>>>
>>> pattern = r'<(?P<tagname>[a-z]+)>.*</(?P=tagname)>'
>>> re.findall(pattern, string)
['p']
>>>
>>> pattern = r'<(?P<tagname>[a-z]+)>(.*)</(?P=tagname)>'
>>> re.findall(pattern, string)
[('p', 'We choose to go to the <strong>Moon</strong>')]

7.23.9. Assignments