Replacing "\"s using re.sub

Andrew M. Kuchling akuchlin at mems-exchange.org
Sat Feb 19 19:14:04 CET 2000


"Edward C. Jones" <edcjones at erols.com> writes:
> I need to replace all instances of "\" in a string with "\\". The
> function "re.sub" looks like the way to do it. "re.sub" is called
> by "re.sub(pattern, repl, string)". For my problem what should
> "pattern" and "repl" be?

If you're only dealing with a substitution of one exact string with
another exact string, it's faster and easier to use string.replace: 

S2 = string.replace(S, '\\', '\\\\')

If you must use regular expressions for some reason (sigh), think
about the problem a bit.  "pattern" needs to match a single \ ; \ is
used as a special character in regular expressions, so to match a
single \ the pattern has to be \\ .

You want the replacement to be \\, but \ is also special in
replacement strings for things like \g<2>, so you need to double them
both.

The patterns need to be inside strings, and Python also uses \ as a
special character; use a raw string, as in r"contents", to prevent
Python from paying attention to \..

S = "String with \\ in it; here's another \\ "
S2 = re.sub(r'\\', r'\\\\', S)
print S
print S2

This outputs:

String with \ in it; here's another \
String with \\ in it; here's another \\

If you didn't use a raw string, you'd need yet another level of \ quoting,
doubling all the backslashes again:

S2 = re.sub('\\\\', '\\\\\\\\', S)

-- 
A.M. Kuchling			http://starship.python.net/crew/amk/
Sending a newgroup message without permission of Leader Kibo: Poster is forced
to adopt twelve wacky sitcom children.
    -- Kibo, in the Happynet Manifesto





More information about the Python-list mailing list