import java.util.regex.Pattern;

public class CharUtil {

    public static void main(String[] args) {
        String[] strArr = new String[] { "www.micmiu.com", "!@#$%^&*()_+{}[]|\"'?/:;<>,.", "?。ぁǎ?；“”‘'《》，。？、", "不要啊", "やめて", "韓佳人", "???" };
        for (String str : strArr) {
            System.out.println("===========> 測(cè)試字符串：" + str);
            System.out.println("正則判斷結(jié)果：" + isChineseByREG(str) + " -- " + isChineseByName(str));
            System.out.println("Unicode判斷結(jié)果 ：" + isChinese(str));
            System.out.println("詳細(xì)判斷列表：");
            char[] ch = str.toCharArray();
            for (int i = 0; i < ch.length; i++) {
                char c = ch[i];
                System.out.println(c + " --> " + (isChinese(c) ? "是" : "否"));
            }
        }
    }

    // 根據(jù)Unicode編碼完美的判斷中文漢字和符號(hào)
    private static boolean isChinese(char c) {
        Character.UnicodeBlock ub = Character.UnicodeBlock.of(c);
        if (ub == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS || ub == Character.UnicodeBlock.CJK_COMPATIBILITY_IDEOGRAPHS
                || ub == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS_EXTENSION_A || ub == Character.UnicodeBlock.CJK_UNIFIED_IDEOGRAPHS_EXTENSION_B
                || ub == Character.UnicodeBlock.CJK_SYMBOLS_AND_PUNCTUATION || ub == Character.UnicodeBlock.HALFWIDTH_AND_FULLWIDTH_FORMS
                || ub == Character.UnicodeBlock.GENERAL_PUNCTUATION) {
            return true;
        }
        return false;
    }

    // 完整的判斷中文漢字和符號(hào)
    public static boolean isChinese(String strName) {
        char[] ch = strName.toCharArray();
        for (int i = 0; i < ch.length; i++) {
            char c = ch[i];
            if (isChinese(c)) {
                return true;
            }
        }
        return false;
    }

    // 只能判斷部分CJK字符（CJK統(tǒng)一漢字）
    public static boolean isChineseByREG(String str) {
        if (str == null) {
            return false;
        }
        Pattern pattern = Pattern.compile("[\\u4E00-\\u9FBF]+");
        return pattern.matcher(str.trim()).find();
    }

    // 只能判斷部分CJK字符（CJK統(tǒng)一漢字）
    public static boolean isChineseByName(String str) {
        if (str == null) {
            return false;
        }
        // 大小寫(xiě)不同：\\p 表示包含，\\P 表示不包含
        // \\p{Cn} 的意思為 Unicode 中未被定義字符的編碼，\\P{Cn} 就表示 Unicode中已經(jīng)被定義字符的編碼
        String reg = "\\p{InCJK Unified Ideographs}&&\\P{Cn}";
        Pattern pattern = Pattern.compile(reg);
        return pattern.matcher(str.trim()).find();
    }
}