How to filter out illegal XML character in Java

Question

I'm building a web service.

Some one put illegal characters into our database.

Now when I try to retrieve those strings and send them via a webservice, the client chokes.

I get an error such as :

com.sun.xml.ws.encoding.soap.DeserializationException: Failed to read a response: javax.xml.bind.UnmarshalException
 - with linked exception:
[com.ctc.wstx.exc.WstxUnexpectedCharException: Illegal character ((CTRL-CHAR, code 18))

How can I remove this character in Java?

I'm looking for quick and dirty. Can I use something like this : stringName.replace('\u0022', ' ') — jeph perro, Jan 20 '11 at 00:50
possible duplicate of [Best way to encode text data for XML in Java?](http://stackoverflow.com/questions/439298/best-way-to-encode-text-data-for-xml-in-java) — Taylor Leese, Jan 20 '11 at 00:51
In the error message, does the `code 18` refer to the illegal character? `0x18 = 24 = control-char "CAN"` or `18 = 0x12 = control-char "DC2"` I've seen the "Illegal character" error before when people managed to get control characters into the database. — Stephen P, Jan 20 '11 at 00:59
I have to assume it is DC2, although I'm not sure how that ever got into our database. — jeph perro, Jan 20 '11 at 17:53
Neat solution with apache Xalan http://stackoverflow.com/a/9635310/489364 — kommradHomer, May 20 '13 at 08:44

score 3 · Answer 1 · answered Jan 20 '11 at 01:13

3

Check this out:

stringName.replaceAll("[^\\p{Print}]", "");

Works like a charm.

answered Jan 20 '11 at 01:13

jeph perro

6,242
26
90
124

score 3 · Accepted Answer · answered Jan 20 '11 at 14:43

/**
 * Function to strip control characters from a string.
 * Any character below a space will be stripped from the string.
 * @param iString the input string to be stripped.
 * @return a string containing the characters from iString minus any control characters.
 */
public String stripControlChars(String iString) {
    StringBuffer result = new StringBuffer(iString);
    int idx = result.length();
    while (idx-- > 0) {
        if (result.charAt(idx) < 0x20 && result.charAt(idx) != 0x9 && 
                result.charAt(idx) != 0xA && result.charAt(idx) != 0xD) {
            if (log.isDebugEnabled()) {
                log.debug("deleted character at: "+idx);
            }
            result.deleteCharAt(idx);
        }
    }
    return result.toString();
}

How to filter out illegal XML character in Java

2 Answers2