crawler_编码转换_unicode(年)

 1 import java.util.regex.Matcher;

 2 import java.util.regex.Pattern;

 3 

 4 /**

 5  * @declare: unicode 帮助类<br>

 6  * @author: cphmvp

 7  * @version: 1.0

 8  * @date: 2014年6月10日上午11:45:34

 9  */

10 public class UnicodeUtils {

11     public static void main(String[] args) {

12         String testStr = "2014&#24180;&#26149;&#22799;&#23395;";

13         System.out.println(getStr(testStr));

14     }

15 

16     /**

17      * @declare:得到可见的字符

18      * @param str

19      *            ：&#24180;

20      * @return 年

21      * @author cphmvp

22      */

23     public static String getStr(String str) {

24         String regex = "&#(\\w{5});";

25         Pattern pa = Pattern.compile(regex);

26         String str0 = str;

27         Matcher matcher = pa.matcher(str);

28         String tmstr0 = null, tmstr = null, zhuanhuanstr = null;

29         while (matcher.find()) {

30             tmstr0 = matcher.group();

31             tmstr = matcher.group(1);

32             // System.out.println(tmstr);

33             zhuanhuanstr = (char) Integer.parseInt(tmstr) + "";

34             str0 = str0.replace(tmstr0, zhuanhuanstr);

35         }

36         return str0;

37     }

38 

39 }