I better idea would be to round off the zip code first and then search it in the XML file. I would recommend Xpath Api for querying the XML file. I very good tutorial about Xpath can be found here
http://www.ibm.com/developerworks/library/x-javaxpathapi/index.html
and as SAXparser itself will look for tags, it won't roundoff them, allover XPath is more prefered for searching XML documents, as it is relatively easy.