Unicode re.sub() doesn't work with \g (group 0)

问题

Why doesn't the \g<0> work with unicode regex?

When I tried to use \g<0> to insert a space before and after the group with normal string regex, it works:

>>> punct = """,.:;!@#$%^&*(){}{}|\/?><"'"""
>>> rx = re.compile('[%s]' % re.escape(punct))
>>> text = '''"anständig"'''
>>> rx.sub(r" \g<0> ",text)
' " anst\xc3\xa4ndig " '
>>> print rx.sub(r" \g<0> ",text)
 " anständig "

but with unicode regex, the space isn't added:

>>> punct = u""",–−—’‘‚”“‟„!£"%$'&)(+*-€/.±°´·¸;:=<?>@§#¡•[˚]»_^`≤…\«¿¨{}|"""
>>> rx = re.compile("["+"".join(punct)+"]", re.UNICODE)
>>> text = """„anständig“"""
>>> rx.sub(ur" \g<0> ", text)
'\xe2\x80\x9eanst\xc3\xa4ndig\xe2\x80\x9c'
>>> print rx.sub(ur" \g<0> ", text)
„anständig“

How do I get \g to work in unicode regex?
If (1) is not possible, how do I get the unicode regex input the space before and after a character in punct?

回答1:

I think you have two errors. First, you are not escaping punct like in the first example with re.escape and you have characters like [] that need to be escaped. And second, text variable is not unicode. Example that works:

>>> punct = re.escape(u""",–−—’‘‚”“‟„!£"%$'&)(+*-€/.±°´·¸;:=<?>@§#¡•[˚]»_^`≤…\«¿¨{}|""")
>>> rx = re.compile("["+"".join(punct)+"]", re.UNICODE)
>>> text = u"""„anständig“"""
>>> print rx.sub(ur" \g<0> ", text)
 „ anständig “

来源：https://stackoverflow.com/questions/19427548/unicode-re-sub-doesnt-work-with-g0-group-0

标签

python