java 正则表达式学习

import java.util.regex.*;

public class Test{  
	public static void main(String args[]) throws Exception{
		//基础
		//String regex = "sales.";
		//String regex = ".a..";
		//String regex = ".a.\\.xls";
		//String regex = "[ns]a.\\.xls";
		//String regex = "[ns]a[0123456789]\\.xls";
		//String regex = "[ns]a[0-9]\\.xls";//同[0123456789]一样
		String regex = "[ns]a[^0-9]\\.xls";
		Test.go(regex, StrOne());

		System.out.println("-----------example two------------------");
		//匹配【】一个字符
		regex = "[Rr]eg[Ee]x";
		Test.go(regex, StrTwo());

		System.out.println("-----------example three------------------");
		//匹配 “+”一个或者多个
		regex = "[\\w.]+@[\\w.]+\\.\\w+";
		Test.go(regex, StrThree());

		System.out.println("-----------example four------------------");
		//匹配 “*”零个或者多个
		regex = "\\w+[\\w.]*@[\\w.]+\\.\\w+";
		Test.go(regex, StrThree());

		System.out.println("-----------example five------------------");
		//匹配 “？”零个或者一个
		regex = "https?://[\\w./]+";
		Test.go(regex, StrFour());

		System.out.println("-----------example five------------------");
		//匹配 “{2,4}”重复出现最少2个，最大4个
		regex = "\\d{1,2}[-\\/]\\d{1,2}[-\\/]\\d{2,4}";
		Test.go(regex, StrFive());

		System.out.println("-----------example six------------------");
		//匹配 “{2,4}”重复出现最少2个，最大4个，{3，}匹配至少3位，{3}匹配3位
		regex = "\\d+:\\$\\d{3,}\\.\\d{2}";
		Test.go(regex, StrSix());

		System.out.println("-----------example seven-----贪婪型-------------");
		/**
		 * 贪婪型		懒惰型
		 * *			*？
		 * +			+？
		 * {n,}			{n,}?
		 */
		regex = "<[Bb]>.*</[Bb]>";
		Test.go(regex, StrSeven());
		System.out.println("-----------example seven-----懒惰型-------------");
		regex = "<[Bb]>.*?</[Bb]>";
		Test.go(regex, StrSeven());
		System.out.println("-----------example eight----\\b--------------");
		//单词边界 \b
		regex = "\\bcat\\b";
		Test.go(regex, StrEight());
		System.out.println("-----------example eight----\\B--------------");
		//不匹配单词边界 \B
		regex = "\\B-\\B";
		Test.go(regex, StrEight());
		System.out.println("-----------example nine-------错误-----------");
		//字符串边界 ^字符串的开头(只有放在【】里边才表示非哦。。。)
		//字符串边界 $结束
		regex = "<\\?xml.*\\?>";
		Test.go(regex, StrNine());
		System.out.println("-----------example nine-------正确（文件开始不是xml）------错误-----");
		regex = "^\\s*<\\?xml.*\\?>";
		Test.go(regex, StrNine());
		System.out.println("-----------example nine-------正确-----------");
		regex = "^\\s*<\\?xml.*\\?>";
		Test.go(regex, StrNine_());
		System.out.println("-----------example nine-------$---错误--------");
		regex = "</[Hh][Tt][Mm][Ll]>\\s*$";
		Test.go(regex, StrNine__());
		System.out.println("-----------example nine-------$---正确--------");
		regex = "</[Hh][Tt][Mm][Ll]>\\s*$";
		Test.go(regex, StrNine___());
		System.out.println("-----------example eleven---------------");
		//字表达式
		regex = "( ){2,}";
		Test.go(regex, StrEleven());
		System.out.println("-----------example twelve---------------");
		//子表达式
		regex = "\\d{1,3}.\\d{1,3}.\\d{1,3}.\\d{1,3}";
		Test.go(regex, StrTwelve());
		regex = "(\\d{1,3}.){3}\\d{1,3}";
		Test.go(regex, StrTwelve());
		System.out.println("-----------example thirteen---------------");
		regex = "(19|20)\\d{2}";
		Test.go(regex, StrThirteen());
		System.out.println("-----------example IP---------------");
		/**
		 * 任何一个1位或2位的数字；
		 * 任何一个以1位开头的3位数字；
		 * 任何一个以2开头，第二位数字在0~4之间的3位数字；
		 * 任何一个以25开头，第三位数字在0~5之间的3位数字；
		 */
		regex = "(((\\d{1,2})|(1\\d{2})|(2[0-4]\\d)|(25[0-5]))\\.){3}((\\d{1,2})|(1\\d{2})|(2[0-4]\\d)|(5[0-5]))";
		Test.go(regex, StrTwelve());
		System.out.println("-----------example fourteen---回溯引用------------");
		//回溯引用，查出重复的
		/**
		 * [ ]匹配一个或者多个空格；
		 * \w+匹配一个或多个字母数字字符；
		 * (\w+)是一个子表达式
		 * 当(\w+)匹配到单词of的时候，\1也匹配单词of
		 * 回溯引用指的是后半部分引用在前半部分中定义的字表达式
		 * \1表示模式里第一个表达式
		 */
		regex = "[ ]+(\\w+)[ ]+\\1";
		Test.go(regex, StrFourteen());
		System.out.println("-----------example Fiveteen-----回溯引用----------");
		regex = "<[hH]([1-6])>.*?</[hH]\\1>";
		Test.go(regex, StrFiveteen());
		System.out.println("-----------example 搜索》替换 给邮件加链接-----回溯引用----------");
		regex = "[\\w.]+@[\\w.]+\\.\\w+";
		Test.go(regex, StrThree());
		regex = StrThree().replaceAll("(\\w+[\\w\\.]*@[\\w\\.]+\\.\\w+)", "<A HREF='mailto:$1'>$1</A>");
		System.out.println(regex);
		System.out.println("-----------example -----向前查找---------");
		//向前查找忽略":" 
		regex = ".+(?=:)";
		Test.go(regex, StrSeventeen());
		regex = ".+(:)";
		Test.go(regex, StrSeventeen());
		System.out.println("-----------example -----向后查找---------");
		//向后查找
		regex = "(?<=\\$)[0-9.]+";
		Test.go(regex, StrEighteen());
		System.out.println("-----------example -----向前向后查找---------");
		//向前向后查找
		regex = "(?<=<[tT][iI][tT][lL][eE]>).*(?=</[tT][iI][tT][lL][eE]>)";
		Test.go(regex, StrNineteen());
	}
	public static String StrNineteen(){
		String str = 
			"<html>" +
			"<title>Hello World</title>" +
			"</html>";
		return str;
	}
	public static String StrEighteen(){
		String str = 
			"ABCD1:$2.45\n" +
			"YHTG8:$56.65";
		return str;
	}
	public static String StrSeventeen(){
		String str = 
			"http://www.google.com/\n" +
			"https://www.sina.com.cn/\n" +
			"htpps://www.wiki.com";
		return str;
	}
	public static String StrSixteen(){
		String str = 
			"313-555-1234\n" +
			"248-555-9999\n" +
			"810-555-9000";
		return str;
	}
	public static String StrFiveteen(){
		String str = 
			"<body>" +
			"<H1>Welcome</H1>" +
			"Hello world hi" +
			"<H2>Welcome</H2>" +
			"Hello world hii" +
			"<H2>Welcome</H3>" +
			"</body>";
		return str;
	}
	public static String StrFourteen(){
		String str = 
			"This is a block of of text,here are are,they and and";
		return str;
	}
	public static String StrThirteen(){
		String str = 
			"ID:0123" +
			"SEX:M" +
			"DOB:1967-08-17" +
			"Status:Avtive";
		return str;
	}
	public static String StrTwelve(){
		String str = 
			"Ping hog. forta.com [12.159.46.200]";
		return str;
	}
	public static String StrEleven(){
		String str = 
			"Hello,my name Windows  2000,";
		return str;
	}
	public static String StrNine___(){
		String str = 
			"<?xml version='1.0' encoding='encoding='UTF-8''?>" +
			"</html>";
		return str;
	}
	public static String StrNine__(){
		String str = 
			"<?xml version='1.0' encoding='encoding='UTF-8''?>" +
			"</html> 1234;";
		return str;
	}
	public static String StrNine_(){
		String str = 
			"<?xml version='1.0' encoding='encoding='UTF-8''?>";
		return str;
	}
	public static String StrNine(){
		String str = 
			"This is bad, real bad!\n" +
			"<?xml version='1.0' encoding='encoding='UTF-8''?>";
		return str;
	}
	public static String StrEight(){
		String str = 
			"The cat cattered his food all over the room.\n" +
			"appears on your color - coded pass-key";
		return str;
	}
	public static String StrSeven(){
		String str = 
			"living in <B>AK</B> and <B>HI</B>";
		return str;
	}
	public static String StrSix(){
		String str = 
			"1001:$496.80\n" +
			"1002:$26.43\n" +
			"1003:$4125.85\n" +
			"1004:$517.98";
		return str;
	}
	public static String StrFive(){
		String str = 
			"4/8/03\n" +
			"10-6-2004\n" +
			"2/2/2\n" +
			"01-01-01";
		return str;
	}
	public static String StrFour(){
		String str = 
			"The URL is http://www.forta.com/,to connect\n" +
			"securely use https://www.forta.com/ insteads\n" +
			"securely use httpssssssss://www.forta.com/";
		return str;
	}
	public static String StrThree(){
		String str = 
			"..ben.forta@forta.com\n" +
			"ben@urgent.forta.com\n" +
			"Hello .ben@fortaben.com is my email addres.";
		return str;
	}
	public static String StrTwo(){
		String str = 
			"The phrase 'regular expression' is often\n" +
			"abbreviated as RegEx or regex.";
		return str;
	}
	public static String StrOne(){
		String str = 
			"sales1.xls\n" +
			"orders3.xls\n"+
			"sales2.xls\n" +
			"sales3.xls\n" +
			"apac1.xls\n" +
			"eroupe2.xls\n" +
			"na1.xls\n" +
			"na2.xls\n" +
			"sa1.xls\n" +
			"ca1.xls\n" +
			"na58xls\n" +
			"sam.xls";
		return str;
	}
	public static void go(String regex,String str){
		Pattern p = Pattern.compile(regex);
		Matcher m = p.matcher(StrOne());
		p = Pattern.compile(regex);
		m = p.matcher(str);
		String val = null; 
		while (m.find()){
			val = m.group();
			System.out.println("MATCH: " + val);
		}
	}
}