Regular expressions in Python and Perl

Python supports essentially the same regular expression syntax as Perl, as far as the regular expressions themselves. However, the syntax for using regular expressions is substantially different.

Regular expression support is not available out of the box; you must import the re module.

Regular expression patterns are contained in strings, in contrast to Perl’s built-in // syntax. This means that some characters need to be escaped in order to be passed on to the regular expression engine. To be safe, always use raw strings (r'' or r"") to contain patterns.

You might think that re.match() is the analog to Perl’s m// match operator. It’s not! The re.match() function matches regular expressions starting at the beginning of a string. It behaves as if every pattern has ^ prepended. The function re.search() behaves like Perl’s  m// and is probably what you want to use exclusively.

The functions match and search return None if no match is found and a match object otherwise. You can retrieve captured matches via the group method on the match object. The group method without any argument returns the entire match. The group method with a positive integer argument returns captured expressions: group(1) returns the first capture, group(2) returns the second, analogous to $1, $2, etc. in Perl.

Python doesn’t have a global modifier like Perl’s /g option. To find all matches to a pattern, use re.findall() rather than re.search(). The findall method returns a list of matches rather than a match object. If the match contains captured subexpressions, findall will return a list of tuples, the tuples being the captures.

To substitute for a pattern, analogous to Perl’s s// operator, use re.sub(). Actually, re.sub() is analogous to s//g since it replaces all instances of a pattern by default. To change this behavior, you can specify the maximum number of instances to replace using the max parameter to re.sub(). Setting this parameter to 1 causes only the first instance to be substituted, as in Perl’s s//.

To make a regular expression case-insensitive, pass the argument re.I (or re.IGNORECASE) as the final argument to re.search().

The function re.sub does not take flags such as re.I. So in order to make the regular expression match case-insensitive, one must modify the regular expression itself by adding (?i) to the beginning of the expression. (The modifier (?i) can go anywhere, but the regular expression will be most readable if the modifier goes at the beginning or possibly at the end.) Also, re.sub does not modify its argument but returns a new string, unlike Perl’s s//.

Resources

Notes on using regular expressions in other languages: PowerShell, C++, R, Mathematica

Other Python articles: making an XML sitemap, languages easy to pick up

For daily tips on regular expressions, follow @RegexTip on Twitter.

Regex tip icon