From: Aaron Patterson <aaron@...>
Date: 2009-08-27T07:48:49+09:00
Subject: Re: Question about Nokogiri xpath

Hi Venkat,

On Thu, Aug 27, 2009 at 06:49:41AM +0900, Venkat Akkineni wrote:
> I am trying to parse an xml file using nokogiri. If any body has
> experience with nokogiri. I request you to help.
> 
> Sample xml file.
> 
> <database name="TABLE AD_CLIENT">
>     <table name="AD_CLIENT" primaryKey="AD_CLIENT_KEY">
>       <column name="AD_CLIENT_ID" primaryKey="true" required="true"
> type="VARCHAR" size="32" autoIncrement="false">
>         <default/>
>         <onCreateDefault/>
>       </column>
> </database>
> 
> It can be seen that Column element has two child elements. But when I
> use
> 
> input = Nokogiri::XML(File.new(file))
> input.xpath('//table/column').each do |column|
> column.children.length
> end
> 
> 
> The result is 5.
> 
> While the <default/> element is at index 1, element <onCreateDefault/>
> is at index 5. Does any one know why this is.

Your XML contains whitespace nodes.  The whitespace nodes are maintained
in the tree that Nokogiri produces.  If you iterate over each node, you
can see this in action:

  input.xpath('//table/column').each do |column|
    # Check each child for blankness
    column.children.each { |child| puts child.blank? }
  end

If you do not care to maintain whitespace nodes, you can tell the parser
to disable blank nodes like this:

  input = Nokogiri::XML(File.new(file)) { |cfg| cfg.noblanks }

I hope that helps!

If you have more questions, I encourage you to join the nokogiri mailing
list which can be found here:

  http://groups.google.com/group/nokogiri-talk

-- 
Aaron Patterson
http://tenderlovemaking.com/