Forum: Ruby Ordering XML Attributes with Hpricot?

Hey,

Is there a way to organize/print out the xml attributes using Hpricot,
or do I have to run through the xml file again and replace patterns?

I would like to be able to say "put this attribute first, put this
attribute next ...", so I can say, I want this:

<node id="name" property="value"/>

not this:

<node property="value" id="name"/>

Since the attributes are kept in a hash there's no order to them, so
they appear in seemingly random order, but it's the same random order
consistently.

Any ideas how to do that?

And is there a way to say "after two attributes, make a new line".  So I
can print out xml that can be edited by humans like code.

Thanks,
Lance

Dear Lance,

> 
> Since the attributes are kept in a hash there's no order to them, so
> they appear in seemingly random order, but it's the same random order
> consistently.
> 
> Any ideas how to do that?
> 

a Hash can be sorted to give an Array with Hash#sort :

http://ruby-doc.org/core/classes/Hash.html#M002865


> And is there a way to say "after two attributes, make a new line".  So I
> can print out xml that can be edited by humans like code.

You can then iterate through the Array with Array#each_with_index,
eg.

my_array.each_with_index{|x,i|

Dear Lance,

I accidentally hit the "send" button too early:

> 
> Since the attributes are kept in a hash there's no order to them, so
> they appear in seemingly random order, but it's the same random order
> consistently.
> 
> Any ideas how to do that?
> 

a Hash can be sorted to give an Array with Hash#sort :

http://ruby-doc.org/core/classes/Hash.html#M002865


> And is there a way to say "after two attributes, make a new line".  So I
> can print out xml that can be edited by humans like code.

You can then iterate through the Array with Array#each_with_index,
eg.

my_array.each_with_index{|x,i| if i%2==0 ; p x + "\n"; else p x; end}

Best regards,

Axel

Thanks a lot axel, I'll give these a try

Best,
Lance

Axel Etzold wrote:
> Dear Lance,
> 
> I accidentally hit the "send" button too early:
> 
>> 
>> Since the attributes are kept in a hash there's no order to them, so
>> they appear in seemingly random order, but it's the same random order
>> consistently.
>> 
>> Any ideas how to do that?
>> 
> 
> a Hash can be sorted to give an Array with Hash#sort :
> 
> http://ruby-doc.org/core/classes/Hash.html#M002865
> 
> 
>> And is there a way to say "after two attributes, make a new line".  So I
>> can print out xml that can be edited by humans like code.
> 
> You can then iterate through the Array with Array#each_with_index,
> eg.
> 
> my_array.each_with_index{|x,i| if i%2==0 ; p x + "\n"; else p x; end}
> 
> Best regards,
> 
> Axel

This means though I have to do two passes on the XML:

1) Modify the nodes with data the way nokogiri or hpricot do it (xpath 
and whatnot)
2) Format the xml using regular expression on pure strings, not using 
the xml parsing engines.

Is that correct?

Thanks,
Lance

-------- Original-Nachricht --------
> Datum: Fri, 25 Sep 2009 10:54:03 +0900
> Von: Lance Pollard <lancejpollard@gmail.com>
> An: ruby-talk@ruby-lang.org
> Betreff: Re: Ordering XML Attributes with Hpricot?

> This means though I have to do two passes on the XML:
> 
> 1) Modify the nodes with data the way nokogiri or hpricot do it (xpath 
> and whatnot)
> 2) Format the xml using regular expression on pure strings, not using 
> the xml parsing engines.
> 
> Is that correct?

Lance,

I remember that Hpricot and Nokogiri both have pretty_print methods, but 
I have never used them. Also, I don't know whether "pretty" can be 
defined so that everybody agrees :)

Best regards,

Axel

2009/9/25 Lance Pollard <lancejpollard@gmail.com>:
> This means though I have to do two passes on the XML:
>
> 1) Modify the nodes with data the way nokogiri or hpricot do it (xpath
> and whatnot)
> 2) Format the xml using regular expression on pure strings, not using
> the xml parsing engines.
>
> Is that correct?

I would not work on the output XML via String replacements.  I would
rather adjust the output process.  For example, if you would be
working with REXML you could implement a Formatter which outputs
attributes in a particular order.  I don't know whether this can be
done with Nokogiri or Hpricot as well or as easily.

Kind regards

robert


require 'rexml/document'

class OrderedAttributes < REXML::Formatters::Pretty
  def write_element(elm, out)
    att = elm.attributes

    class <<att
      alias _each_attribute each_attribute

      def each_attribute(&b)
        to_enum(:_each_attribute).sort_by {|x| x.name}.each(&b)
      end
    end

    super(elm, out)
  end
end

doc = REXML::Document.new(DATA.read)

fmt = REXML::Formatters::Pretty.new
fmt.write(doc, $stdout)
puts

fmt = OrderedAttributes.new
fmt.write(doc, $stdout)
puts

__END__
<foo battr="1" aattr="2" cattr="3">
</foo>

 > require 'rexml/document'
> 
> class OrderedAttributes < REXML::Formatters::Pretty
>   def write_element(elm, out)
>     att = elm.attributes
> 
>     class <<att
>       alias _each_attribute each_attribute
> 
>       def each_attribute(&b)
>         to_enum(:_each_attribute).sort_by {|x| x.name}.each(&b)
>       end
>     end
> 
>     super(elm, out)
>   end
> end
> 
> doc = REXML::Document.new(DATA.read)
> 
> fmt = REXML::Formatters::Pretty.new
> fmt.write(doc, $stdout)
> puts
> 
> fmt = OrderedAttributes.new
> fmt.write(doc, $stdout)
> puts
> 
> __END__
> <foo battr="1" aattr="2" cattr="3">
> </foo>

Thanks a lot Robert, I will try that out immediately.

Best,
Lance