java 判断字符串是否为乱码

妖狐艹你老母 2022-05-09 18:44 257阅读 0赞

以下是一个事例：

import java.util.regex.Matcher;
    import java.util.regex.Pattern;
    
    public class MessyCodeCheck {
    
    	public static boolean isChinese(char c) {
    		Character.UnicodeBlock ub = Character.UnicodeBlock.of(c);
    		return ub == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS
    				|| ub == Character.UnicodeBlock.CJK_COMPATIBILITY_IDEOGRAPHS
    				|| ub == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS_EXTENSION_A
    				|| ub == Character.UnicodeBlock.GENERAL_PUNCTUATION
    				|| ub == Character.UnicodeBlock.CJK_SYMBOLS_AND_PUNCTUATION
    				|| ub == Character.UnicodeBlock.HALFWIDTH_AND_FULLWIDTH_FORMS;
    	}
    
    	/**
    	 *判断字符串是否为乱码
    	 * @param strName
    	 * @return
    	 */
    	public static boolean isMessyCode(String strName) {
    		Pattern p = Pattern.compile("\\s*|\t*|\r*|\n*");
    		Matcher m = p.matcher(strName);
    		String after = m.replaceAll("");
    		String temp = after.replaceAll("\\p{P}", "");
    		char[] ch = temp.trim().toCharArray();
    		float chLength = ch.length;
    		float count = 0;
    		for (char c : ch) {
    			if (!Character.isLetterOrDigit(c)) {
    				if (!isChinese(c)) {
    					count = count + 1;
    					System.out.print(c);
    				}
    			}
    		}
    		float result = count / chLength;
    		return result > 0.4;
    
    	}
    
    	public static void main(String[] args) {
    		System.out.println(isMessyCode("*��JTP.jar�ļ����JTP�ļ���ȡ��ͼƬ��Դ"));
    		System.out.println(isMessyCode("你好"));
    		System.out.println(isMessyCode("123123adfasdfsa"));
    		System.out.println(isMessyCode("123123a测试——@#￥%%dfasdfsa"));
    
    		System.out.println(isMessyCode("*��JTP.jar�ļ����JTP�ļ���ȡ��ͼƬ��Դ"));
    		System.out.println(isMessyCode("你好"));
    		System.out.println(isMessyCode("123123adfasdfsa"));
    		System.out.println(isMessyCode("123123a测试——@#￥%%dfasdfsa"));
    	}
    }