Ruby bindings to re2, an "efficient, principled regular expression library".
C++ Ruby
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
Gemfile.lock Restore Gemfile.lock for Travis CI Sep 11, 2016

re2 Build Status

A Ruby binding to re2, an "efficient, principled regular expression library".

Current version: 1.0.0
Supported Ruby versions: 1.8.7, 1.9.2, 1.9.3, 2.0.0, 2.1.0, 2.2, 2.3, Rubinius 2.2


You will need re2 installed as well as a C++ compiler such as gcc (on Debian and Ubuntu, this is provided by the build-essential package). If you are using Mac OS X, I recommend installing re2 with Homebrew by running the following:

$ brew install re2

If you are using Debian, you can install the libre2-dev package like so:

$ sudo apt-get install libre2-dev

Recent versions of re2 require a compiler with C++11 support such as clang 3.4 or gcc 4.8.

If you are using a packaged Ruby distribution, make sure you also have the Ruby header files installed such as those provided by the ruby-dev package on Debian and Ubuntu.

You can then install the library via RubyGems with gem install re2 or gem install re2 -- --with-re2-dir=/opt/local/re2 if re2 is not installed in the default location of /usr/local/.


Full documentation automatically generated from the latest version is available at

Note that re2's regular expression syntax differs from PCRE and Ruby's built-in Regexp library, see the official syntax page for more details.


While re2 uses the same naming scheme as Ruby's built-in regular expression library (with Regexp and MatchData), its API is slightly different:

$ irb -rubygems
> require 're2'
> r ='w(\d)(\d+)')
=> #<RE2::Regexp /w(\d)(\d+)/>
> m = r.match("w1234")
=> #<RE2::MatchData "w1234" 1:"1" 2:"234">
> m[1]
=> "1"
> m.string
=> "w1234"
> m.begin(1)
=> 1
> m.end(1)
=> 2
> r =~ "w1234"
=> true
> r !~ "bob"
=> true
> r.match("bob")
=> nil

As (or RE2::Regexp.compile) can be quite verbose, a helper method has been defined against Kernel so you can use a shorter version to create regular expressions:

> RE2('(\d+)')
=> #<RE2::Regexp /(\d+)/>

Note the use of single quotes as double quotes will interpret \d as d as in the following example:

> RE2("(\d+)")
=> #<RE2::Regexp /(d+)/>

As of 0.3.0, you can use named groups:

> r ='(?P<name>\w+) (?P<age>\d+)')
=> #<RE2::Regexp /(?P<name>\w+) (?P<age>\d+)/>
> m = r.match("Bob 40")
=> #<RE2::MatchData "Bob 40" 1:"Bob" 2:"40">
> m[:name]
=> "Bob"
> m["age"]
=> "40"

As of 0.6.0, you can use RE2::Regexp#scan to incrementally scan text for matches (similar in purpose to Ruby's String#scan). Calling scan will return an RE2::Scanner which is enumerable meaning you can use each to iterate through the matches (and even use Enumerator::Lazy):

re = RE2('(\w+)')
scanner = re.scan("It is a truth universally acknowledged")
scanner.each do |match|
  puts match


enum = scanner.to_enum #=> ["It"] #=> ["is"]


  • Pre-compiling regular expressions with, RE2::Regexp.compile(re) or RE2(re) (including specifying options, e.g."pattern", :case_sensitive => false)

  • Extracting matches with re2.match(text) (and an exact number of matches with re2.match(text, number_of_matches) such as re2.match("123-234", 2))

  • Extracting matches by name (both with strings and symbols)

  • Checking for matches with re2 =~ text, re2 === text (for use in case statements) and re2 !~ text

  • Incrementally scanning text with re2.scan(text)

  • Checking regular expression compilation with re2.ok?, re2.error and re2.error_arg

  • Checking regular expression "cost" with re2.program_size

  • Checking the options for an expression with re2.options or individually with re2.case_sensitive?

  • Performing a single string replacement with pattern.replace(replacement, original)

  • Performing a global string replacement with pattern.replace_all(replacement, original)

  • Escaping regular expressions with RE2.escape(unquoted) and RE2.quote(unquoted)


  • Thanks to Jason Woods who contributed the original implementations of RE2::MatchData#begin and RE2::MatchData#end;
  • Thanks to Stefano Rivera who first contributed C++11 support.


All feedback should go to the mailing list: