正则表达式

Posted 2020-06-23

//自己看教程写的代码，感觉涵盖了正则大部分的知识点，基本够用于平常工作（一些地方的注释写给自己看的，可以略过:D）

package basic;

import java.io.BufferedReader;
import java.io.FileReader;
import java.io.IOException;
import java.io.InputStreamReader;
import java.net.URL;
import java.net.URLConnection;
import java.util.Arrays;
import java.util.Set;
import java.util.TreeSet;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

/**
 * 正则表达式
 * <ul>
 * <li>\d (digital) \D (none digital)</li>
 * <li>\s (space) \S (none space)</li>
 * <li>\w (word) \W (none word)</li>
 * <li>. (anyone)</li>
 * </ul>
 * <ul>
 * <li>? (zero or one)</li>
 * <li>+ (one or more)</li>
 * <li>* (zero or more)</li>
 * <li>{n,m} (n<= and =<m)</li>
 * </ul>
 * <ul>
 * 后向引用见splitDemo() matcherDemo() demo1() demo2()
 * <li>content.replaceAll("<courseId:([0-9]+)>", "<cms:course st=‘name‘ uid=‘$1‘/>")(<换成&lt;可以正常显示)</li>
 * </ul>
 * @author Tank
 * @date 2016年3月10日
 */
public class RegexDemo {
    
    private static String EMAIL_REGEX = "[a-zA-Z0-9_][email protected][a-zA-Z0-9]+(\\.[a-zA-Z]+){1,3}";

    public static void main(String[] args) throws IOException {
//        System.out.println(isQQ("0124309"));
//        System.out.println(isEmail("[email protected]"));
//        System.out.println(isTel("1321234123a"));
//        splitDemo();
//        replaceDemo();
//        matcherDemo();
//        demo1();
//        demo2();
        demo3();
    }
    
    /**
     * 校验QQ号
     * <br>
     * 5~15位，0不能开头，只能是数字
     * 
     * @param s
     * @return boolean
     */
    public static boolean isQQ(String str){
//        String regex = "[1-9][0-9]{4,14}";
        String regex = "[1-9]\\d{4,14}";
        return str.matches(regex);
    }
    
    /**
     * 校验邮箱
     * <br>
     * (大小写字母和数字)@(小写字母).(小写字母和数字).(小写字母)
     * 
     * @param str
     * @return
     */
    public static boolean isEmail(String str){
        String regex = EMAIL_REGEX;//精确匹配
        return str.matches(regex);
    }
    
    /**
     * 校验手机号
     * <br>
     * 13x xxxx xxxx, 15x xxxx xxxx, 18x xxxx xxxx
     * 
     * @param str
     * @return
     */
    public static boolean isTel(String str){
        String regex = "1[358]\\d{9}";
        return str.matches(regex);
    }
    
    /**
     * 按多空格进行切割
     * 
     */
    public static void splitDemo(){
        /*
         * 多空格切割
        String str = "123asdf asuz   asdfwq  asfq2asdf s asdf asf   asdf ds";
        String regex = " +";
         */
        /*
         * 小数点切割
        String str = "asdfasd.zdfasd.asdf.asdfzzz.adfa.a.a.aa.as";
        String regex = "\\.";
        // .表示任意字符 \.表示字符. \\.代码反斜杠转义
        */
        /*
         * 反斜杠切割
        String str = "c:\\abc\\a.txt";
        String regex = "\\\\";
        */
        /*
         * 叠词切割
         */
        String str = "asdffasddfzxxxc";
        String regex = "(.)\\1+";
        //引用
        String[] ss = str.split(regex);
        System.out.println(Arrays.toString(ss));
    }
    
    /**
     * 将叠词替换成单个字母（后向引用）
     * 
     */
    public static void replaceDemo(){
        String str = "asdffasddfzxxxc";
        String regex = "(.)\\1+";
        String newStr = "$1";
        str = str.replaceAll(regex, newStr);
        System.out.println(str);
    }
    
    /**
     * 将文本中三个字母的单词返回
     * <br>
     * 正则对象，匹配器，匹配查找
     * <br>
     * \b 单词边界
     */
    public static void matcherDemo(){
        String str = "adsf sdf zzxjaj. zjd jsh saa, sdf asdf(lei)";
        String regex = "\\b\\w{3}\\b";
        Pattern p = Pattern.compile(regex);
        //将正则封装成对象
        Matcher m = p.matcher(str);
        //获取匹配器对象(将正则对象与字符串关联)
        while(m.find()){
            //将匹配器作用于字符串上，执行一次查找
            System.out.println(m.group());
            //再执行一次返回匹配查找结果
            System.out.println(m.start()+"..."+m.end());
        }
    }
    
    /**
     * 将下列字符串转换成：我要学编程
     * 
     */
    public static void demo1(){
        String str = "我我...我我我...要要要...学学......编....编编编.....程程程....程程";
        str = str.replaceAll("\\.","");
        str = str.replaceAll("(.)\\1+", "$1");
        System.out.println(str);
    }
    
    /**
     * 将IP地址进行地址段顺序的排序
     * 
     */
    public static void demo2(){
        String str = "192.68.1.254 102.49.23.13 10.10.10.10 10.101.88.95 2.2.2.2 8.109.90.30";
        //一位的补一个0，两位的补两个0
        /*
        str = str.replaceAll("(\\b\\d{1}\\b)", "00$1");
        str = str.replaceAll("(\\b\\d{2}\\b)", "0$1");
        */
        //先统一补两个零（零的个数等于IP地址分段长度减1），再截取每个分段后三位数字
        str = str.replaceAll("(\\d+)", "00$1");
        str = str.replaceAll("0*(\\d{3})","$1");
        
        System.out.println(str);
        String regex = "(\\d{3}\\.){3}\\d{3}";
        Pattern p = Pattern.compile(regex);
        Matcher m = p.matcher(str);
        Set<String> ss = new TreeSet<String>();
        while(m.find()){
            ss.add(m.group());
        }
        for (String string : ss) {
            String replaceAll = string.replaceAll("0*(\\d+)","$1");
            System.out.println(replaceAll);
        }
    }
    
    /**
     * 网页爬虫
     * <br>
     * 获取一个网页文件中所有的邮箱地址
     * @throws IOException 
     * 
     */
    public static void demo3() throws IOException{
        /*
        URL url = new URL("http://www.baidu.com");
        URLConnection conn = url.openConnection();
        BufferedReader urlBr = new BufferedReader(new InputStreamReader(conn.getInputStream()));
         */
        //以下是单机示例
        BufferedReader br = new BufferedReader(new FileReader(IODemo.ROOT_PATH+"mail.txt"));
        String line = null;
        Pattern p = Pattern.compile(EMAIL_REGEX);
        while( (line=br.readLine()) !=null){
            Matcher m = p.matcher(line);
            while(m.find()){
                System.out.println(m.group());
            }
        }
        br.close();
    }
}