REXML, a Ruby-style XML toolkit

What’s the difference between results from code (1) and (2) below ?
(element is an XML element)

Code (1), use Element#elements :

element.elements.each do |e|
 puts e.inspect
end

Code (2), use Element#to_a :

element.to_a.each do |e|
 puts e.inspect
end

Update: We can actually use just element.each .. no .to_a requied — thanks to P’Pok for this

Code (2) will give us texts, elements (as well as other nodes).
Where code (1) will give us only elements.

If our input is:

<p><b>bold</b> text</p>

Code (1) will give:

<b> ... </>

While code (2) will give:

<b> ... </>
" text"

This tiny difference already wasted me hours, shamed 🙁
I was thought that text is a kind of element, … that’s plain wrong,
both text and element are kinds of node !

For several REXML tutorials/examples I’ve found, where I copied and pasted codes from for my quick-n-dirty-self-education, all of them show only the use of Element#elements but not Element#to_a.

This is probably because all of them only deal with a data-oriented XML, where a use of ‘mix content’ is rare (and indeed not recommended).
But that’s no longer true for document/text-oriented XML — for example, XHTML.

If you going to process a XML with mix content, beware of #elements.

Correct me if I do anything wrong here.
REXML veteran? Share! 😉

Tutorials:
REXML Home |
XML.com |
developerWorks
API docs