diff a ruby string or array

StackOverflow https://stackoverflow.com/questions/80091

  •  09-06-2019
  •  | 
  •  

Question

How do I do a diff of two strings or arrays in Ruby?

Was it helpful?

Solution

diff.rb is what you want, which is available at http://users.cybercity.dk/~dsl8950/ruby/diff.html via internet archive:

http://web.archive.org/web/20140421214841/http://users.cybercity.dk:80/~dsl8950/ruby/diff.html

OTHER TIPS

For arrays, use the minus operator. For example:

>> foo = [1, 2, 3]
=> [1, 2, 3]
>> goo = [2, 3, 4]
=> [2, 3, 4]
>> foo - goo
=> [1]

Here the last line removes everything from foo that is also in goo, leaving just the element 1. I don't know how to do this for two strings, but until somebody who knows posts about it, you could just convert each string to an array, use the minus operator, and then convert the result back.

I got frustrated with the lack of a good library for this in ruby, so I wrote http://github.com/samg/diffy. It uses diff under the covers, and focuses on being convenient, and providing pretty output options.

For strings, I would first try out the Ruby Gem that @sam-saffron mentioned below. It's easier to install: http://github.com/pvande/differ/tree/master

gem install differ

irb
require 'differ'

one = "one two three"
two = "one two 3"

Differ.format = :color
puts Differ.diff_by_word(one, two).to_s

Differ.format = :html
puts Differ.diff_by_word(one, two).to_s

The HTMLDiff that @da01 mentions above worked for me. 

script/plugin install git://github.com/myobie/htmldiff.git

# bottom of environment.rb
require 'htmldiff'

# in model
class Page < ActiveRecord::Base
  extend HTMLDiff
end

# in view
<h1>Revisions for <%= @page.name %></h1>
<ul>
<% @page.revisions.each do |revision| %>
  <li>
    <b>Revised <%= distance_of_time_in_words_to_now revision.created_at %> ago</b><BR>
      <%= Page.diff(
        revision.changes['description'][0],
        revision.changes['description'][1]
      ) %>
      <BR><BR>
  </li>
<% end %>

# in style.css
ins.diffmod, ins.diffins { background: #d4fdd5; text-decoration: none; }
del.diffmod, del.diffdel { color: #ff9999; }

Looks pretty good. By the way I used this with the acts_as_audited plugin.

There is also diff-lcs which is available as a gem. It hasn't been updated since 2004 but we have been using it without any problem.

Edit: A new version was released in 2011. Looks like it's back in active development.

http://rubygems.org/gems/diff-lcs

t=s2.chars; s1.chars.map{|c| c == t.shift ? c : '^'}.join

This simple line gives a ^ in the positions that don't match. That's often enough and it's copy/paste-able.

I just found a new project that seems pretty flexible:

http://github.com/pvande/differ/tree/master

Trying it out and will try to post some sort of report.

I had the same doubt and the solution I found is not 100% ruby, but is the best for me. The problem with diff.rb is that it doesn't have a pretty formatter, to show the diffs in a humanized way. So I used diff from the OS with this code:

 def diff str1, str2
   system "diff #{file_for str1} #{file_for str2}"
 end

 private
 def file_for text
   exp = Tempfile.new("bk", "/tmp").open
   exp.write(text)
   exp.close
   exp.path
 end

Just for the benefit of Windows people: diffy looks brilliant but I belive it will only work on *nix (correct me if I'm wrong). Certainly it didn't work on my machine.

Differ worked a treat for me (Windows 7 x64, Ruby 1.8.7).

To get character by character resolution I added a new function to damerau-levenshtein gem

require "damerau-levenshtein"
differ = DamerauLevenshtein::Differ.new
differ.run "Something", "Smothing"
# returns ["S<ins>o</ins>m<subst>e</subst>thing", 
#  "S<del>o</del>m<subst>o</subst>thing"]

or with parsing:

require "damerau-levenshtein"
require "nokogiri"

differ = DamerauLevenshtein::Differ.new
res = differ.run("Something", "Smothing!")
nodes = Nokogiri::XML("<root>#{res.first}</root>")

markup = nodes.root.children.map do |n|
  case n.name
  when "text"
    n.text
  when "del"
    "~~#{n.children.first.text}~~"
  when "ins"
    "*#{n.children.first.text}*"
  when "subst"
    "**#{n.children.first.text}**"
  end
end.join("")

puts markup
Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top