Filtering illegal XML characters in Java

Question

XML spec defines a subset of Unicode characters which are allowed in XML documents: http://www.w3.org/TR/REC-xml/#charsets.

How do I filter out these characters from a String in Java?

simple test case:

  Assert.equals("", filterIllegalXML(""+Character.valueOf((char) 2)))

Answer 1

It's not trivial to find out all the invalid chars for XML. You need to call or reimplement the XMLChar.isInvalid() from Xerces,

http://kickjava.com/src/org/apache/xerces/util/XMLChar.java.htm

Answer 2

This page includes a Java method for stripping out invalid XML characters by testing whether each character is within spec, though it doesn't check for highly discouraged characters

Incidentally, escaping the characters is not a solution since the XML 1.0 and 1.1 specs do not allow the invalid characters in escaped form either.