Java过滤Unicode

我们在解析XML文件时，会碰到程序发生以下一些异常信息：

引用
org.xml.sax.SAXParseException: An invalid XML character (Unicode: 0x{2}) was found in the value of attribute "{1}" and element is "1f".

引用
An invalid XML character (Unicode: 0x1d) was found in the CDATA section.

这些错误的发生是由于一些不可见的特殊字符的存在，而这些字符对于XMl文件来说又是非法的，所以XML解析器在解析时会发生异常，官方定义了XML的无效字符分为三段：

0x00 - 0x08
0x0b - 0x0c
0x0e - 0x1f



解决方法是：在解析之前先把字符串中的这些非法字符过滤掉：


string.replaceAll("[\\x00-\\x08\\x0b-\\x0c\\x0e-\\x1f]", "")



测试代码：TestXmlInvalidChar.java


package michael.xml;

import java.io.ByteArrayInputStream;

import javax.xml.parsers.DocumentBuilder;
import javax.xml.parsers.DocumentBuilderFactory;

import org.w3c.dom.Document;
import org.w3c.dom.Element;

/**
* @author michael
*
*/
public class TestXmlInvalidChar {

/**
* @param args
*/
public static void main(String[] args) {

// 测试的字符串应该为：<r><c d="s" n="j"></c></r>
// 正常的对应的byte数组为
byte[] ba1 = new byte[] { 60, 114, 62, 60, 99, 32, 100, 61, 34, 115,
34, 32, 110, 61, 34, 106, 34, 62, 60, 47, 99, 62, 60, 47, 114,
62 };
System.out.println("ba1 length=" + ba1.length);
String ba1str = new String(ba1);
System.out.println(ba1str);
System.out.println("ba1str length=" + ba1str.length());
System.out.println("-----------------------------------------");
// 和正常的byte 数组相比 多了一个不可见的 31
byte[] ba2 = new byte[] { 60, 114, 62, 60, 99, 32, 100, 61, 34, 115,
34, 32, 110, 61, 34, 106, 31, 34, 62, 60, 47, 99, 62, 60, 47,
114, 62 };
System.out.println("ba2 length=" + ba2.length);
String ba2str = new String(ba2);
System.out.println(ba2str);
System.out.println("ba2str length=" + ba2str.length());
System.out.println("-----------------------------------------");
try {
DocumentBuilderFactory dbfactory = DocumentBuilderFactory
.newInstance();
dbfactory.setIgnoringComments(true);
DocumentBuilder docBuilder = dbfactory.newDocumentBuilder();

// 过滤掉非法不可见字符 如果不过滤 XML解析就报异常
String filter = ba2str.replaceAll(
"[\\x00-\\x08\\x0b-\\x0c\\x0e-\\x1f]", "");
System.out.println("过滤后的length=" + filter.length());
ByteArrayInputStream bais = new ByteArrayInputStream(filter
.getBytes());
Document doc = docBuilder.parse(bais);
Element rootEl = doc.getDocumentElement();
System.out.println("过滤后解析正常 root child length="
+ rootEl.getChildNodes().getLength());
} catch (Exception e) {
e.printStackTrace();
}
}

}


测试代码运行结果如下：
引用

ba1 length=26
<r><c d="s" n="j"></c></r>
ba1str length=26
-----------------------------------------
ba2 length=27
<r><c d="s" n="j"></c></r>
ba2str length=27
-----------------------------------------
过滤后的length=26
过滤后解析正常 root child length=1

对比可见，byte数组及字符串的长度前后是不一样的，但打印到控制台显示的结果却是一样的。同样过滤之后的字符串长度是有变化的。


参考：http://sjsky.iteye.com/blog/1055063
http://www.blogjava.net/fingki/archive/2008/09/04/226969.html

--复旦检索 图书馆报错：
org.xml.sax.SAXParseException: An invalid XML character (Unicode: 0x{2}) was found in the value of attribute "{1}" and element is "1f".

String xmlCode2 = HttpClientUtil.getWebInfoByHttpClientGetMethodGBK(searchURL); // 抓取网页
xmlCode2 = xmlCode2.replaceAll("[\\x00-\\x08\\x0b-\\x0c\\x0e-\\x1f]", "");//过滤Unicode