Given a piece of XML:

<Export>
  <Product>
    <SKU>403276</SKU>
    <ItemName>Trivet</ItemName>
    <CollectionNo>0</CollectionNo>
    <Pages>0</Pages>
  </Product>
</Export>

One might assume that REXML is the way to parse it, but we all know how slow it is.

Enter _why’s HTML parser, Hpricot. It’s written in C and since XHTML is a subset of XML, there’s no reason it shouldn’t be able to parse my file.

Turns out it does, it’s really fast, and the code is dead simple.

FIELDS = %w[SKU ItemName CollectionNo Pages]

doc = Hpricot.parse(File.read("my.xml"))
(doc/:product).each do |xml_product|
  product = Product.new
  for field in FIELDS
    product[field] = (xml_product/field.intern).first.innerHTML
  end
  product.save
end

Update: Slight refactoring of the code above. Chris figured out last night that you can use innerHTML which eliminated the only ugly part of the code.