You have a list of the <book> elements already. And you iterate through that list. So in my mind, the real question is: why don't you just get the "isbn" attribute of those <book> elements? I don't see the point of running another XPath expression to get the value of an attribute, even if you were using the correct XPath expression.
Surfs up space ponies, I'm making gravy without this lumpy, tiny ad:
Gift giving made easy with the permaculture playing cards