Question

So im trying to parse and xml file in android. Bellow is a link to the file and my code. I’m trying to get the current temp stored as “temp_f” in the “current_conditions” element. But every time i run it i get this error, 09-15 20:26:47.359: DEBUG/ErOr(17663): java.lang.ClassCastException: org.apache.harmony.xml.dom.ElementImpl

http://www.google.com/ig/api?weather=10598

DocumentBuilderFactory factory1 = DocumentBuilderFactory.newInstance();
            factory1.setNamespaceAware(true); // never forget this!
            InputSource source = new InputSource(new StringReader(xmltemp));
            Document doc = factory1.newDocumentBuilder().parse(source);

            String newstring = "" + ((DocumentBuilderFactory) ((Document) doc.getDocumentElement().
               getElementsByTagName("current_conditions").item(0)).
               getElementsByTagName("temp_f").item(0)).
               getAttribute("data");

Answer 1

Have a look at the Java XPath API — that should be much nicer than traversing the DOM to find your data:

URL url = new URL("http://www.google.com/ig/api?weather=10598");
InputSource xml = new InputSource(url.openStream());
XPath xpath = XPathFactory.newInstance().newXPath();
String data = xpath.evaluate("//current_conditions/temp_f/@data", xml);

---

As for the original question:

First put every statement of your extraction code on a line of its own, this will help you a lot with finding the problems (and line numbers in stacktrace will be a lot more precise):

Element docElem = doc.getDocumentElement();
Document curr_cond = (Document) /*1*/ docElem.getElementsByTagName("current_conditions").item(0);
DocumentBuilderFactory temp_f = (DocumentBuilderFactory) /*2*/ curr_cond.getElementsByTagName("temp_f").item(0);
String newstring = "" + temp_f.getAttribute("data");

At /*1*/ you try to cast a org.w3.dom.Node into an org.w3.dom.Document — but not every Node is a Document, as can be seen by the ClassCastException here: The returned Node is actually an instance of the org.w3.dom.Element interface, more precisely an instance of class of org.apache.harmony.xml.dom.ElementImpl (part of the Apache Harmony XML parser implementation).

So let’s remove the cast and change the declared type of curr_cond to Node:

Element docElem = doc.getDocumentElement();
Node curr_cond = docElem.getElementsByTagName("current_conditions").item(0);
DocumentBuilderFactory temp_f = (DocumentBuilderFactory) /*2*/ curr_cond.getElementsByTagName("temp_f").item(0);
String newstring = "" + temp_f.getAttribute("data");

Now a second problem becomes obvious at /*2*/: The Node interface does not specify a getElementsByTagName() method, that’s probably why you did try to cast to Document above. This won’t compile and there is no way to get it to.

Why you cast the result to DocumentBuilderFactory is beyond me. Yes, there is a getAttribute() method but both the class name and method javadoc should have made clear that it’s definitely not doing want you want.

So my above recommendation holds: Please use XPath, it’s the better tool for this task 🙂

---

There actually is a way to extract the data using only DOM for your first task (current temperature). It relies on the fact that the temp_f element is unique and is a lot more verbose than the above XPath expression, though:

Element docElem = doc.getDocumentElement();
Node temp_f = docElem.getElementsByTagName("temp_f").item(0);
NamedNodeMap attributes = temp_f.getAttributes();
Node dataAttr = attributes.getNamedItem("data");
String data = dataAttr.getNodeValue();