Repairing broken XML file - removing extra less-than/greater-than signs

Question

I have a large XML file which in the middle contains the following:

<ArticleName>Article 1 <START  </ArticleName>

Obviously libxml and other XML libraries can't read this because the less-than sign opens a new tag which is never closed. My question is, is there anything I can do to fix issues like this automatically (preferably in Ruby)? The solution should of course work for any field which has an error like this. Someone said SAX parsing could do the trick but I'm not sure how that would work.

Answer 1

You could do a regular expression search-and-replace, looking for <(?=[^<>]*<) and replacing with <.

In Ruby,

result = subject.gsub(/<(?=[^<>]*<)/, '&lt;')

The rationale behind this being that you want to find < that don't have a corresponding >. Therefore, the regex only matches a < if it is followed by another without any > in-between.

EDIT: Improved the regex by using lookahead. I first thought Ruby didn't support lookahead, but it does. Just not lookbehind...

Answer 2

Nokogiri supports some options for handling bad XML. These might help:

http://rubyforge.org/pipermail/nokogiri-talk/2009-February/000066.html http://nokogiri.org/tutorials/ensuring_well_formed_markup.html

I just messed around with the broken fragment and Nokogiri handles it very nicely:

#!/usr/bin/ruby

require 'rubygems'
require 'nokogiri'

doc = Nokogiri::XML('<?xml version="1.0"?><ArticleName>Article 1 <START  </ArticleName></xml>')
doc.to_s  # => "<?xml version=\"1.0\"?>\n<ArticleName>Article 1 <START/></ArticleName>\n"
doc.errors # => [#<Nokogiri::XML::SyntaxError: error parsing attribute name