0%

Java 正则表达式(二):Java API

1. Java API 中正则表达式相关的类有

  • 正则表达式相关的类位于包 java.util.regex 下,有两个主要的类 PatternMatcher

    • Pattern:表示正则表达式对象,它与要处理的具体字符串无关
    • Matcher:表示一个匹配,它将正则表达式应用于一个具体字符串,通过它对字符串进行处理
  • 正则表达式在 Java 中是需要先以字符串形式表示的

2. 在 Java 中怎样表示正则表达式

  • 在 Java 中,没有什么特殊语法能直接表示正则表达式,正则表达式需要用字符串表示

    • 而在字符串中,\ 也是一个元字符,为了在字符串中表示正则表达式的 \,就需要使用两个 \,即 \\
    • 要匹配 \ 本身,就需要 4 个 \,即 \\\\。比如,表达式 <(\w+)>(.*)</\1> 对应的字符串表示就是 <(\\w+)>(.*)</\\1>
    • 一个简单的规则是:正则表达式中的任何一个 \,在字符串中,需要替换为两个 \
  • 字符串表示的正则表达式可以被编译为一个 Pattern 对象。比如 String regex = "<(\\w+)>(.*)</\\1>";,可以被编译为 Pattern pattern = Pattern.compile(regex);

    • Pattern 是正则表达式的面向对象表示。所谓编译,简单理解就是将字符串表示为了一个内部结构,这个结构是一个有穷自动机
    • 编译有一定的成本,而且 Pattern 对象只与正则表达式有关,与要处理的具体文本无关,它可以安全地被多线程共享。所以,在使用同一个正则表达式处理多个文本时,应该尽量重用同一个 Pattern 对象,避免重复编译
  • Patterncompile() 方法接受一个额外参数,可以指定匹配模式public static Pattern compile(String regex, int flags)。其中,单行模式(点号模式)、多行模式和大小写无关模式,对应的常量分别是:Pattern.DOTALLPattern.MULTILINEPattern.CASE_INSENSITIVE。多个模式可以一起使用,通过 | 连起来即可,比如:Pattern.compile(regex, Pattern.CASE_INSENSITIVE | Pattern.DOTALL);

    • 还有一个模式 Pattern.LITERAL,在此模式下,正则表达式字符串中的元字符将失去特殊含义,被看作普通字符
    • Pattern 有一个静态方法:public static String quote(String s),与模式 Pattern.LITERAL 类似,它将 s 中的字符都看作普通字符。quote() 方法基本上就是在字符串 s 的前后加了 \Q\E。比如,如果 s\\d{6},则 quote() 的返回值就是 \\Q\\d{6}\\E

3. 怎样理解 Stringsplit() 方法:public String[] split(String regex)

  • split() 方法将参数 regex 看作正则表达式,而不是普通的字符。所以,如果分隔符是元字符,比如 . $ | () [ { ^ ? * + \,就需要转义。比如,按点号 . 分隔,需要写为:String[] fields = str.split("\\.");

    • 如果分隔符是用户指定的,程序事先不知道,可以通过 Pattern.quote() 将其看作普通字符串
  • 既然是正则表达式,分隔符就不一定是一个字符。比如,可以将一个或多个空白字符或点号作为分隔符,如下:String str = "abc def hello.\n world";,String [] fields = str.split("[\\s.]+");fields 内容是 [abc, def, hello, world]

    • 需要说明的是,尾部的空白字符串不会包含在返回的结果数组中,但头部和中间的空白字符串会被包含在内
    • 如果字符串中找不到匹配 regex 的分隔符,返回数组长度为 1,元素为原字符串

4. Patternsplit() 方法:public String[] split(CharSequence)Stringsplit() 方法:public String[] split(String regex) 的区别是

  • Pattern 接受的参数是 CharSequence,更为通用。String StringBuilder StringBuffer CharBuffer 等都实现了该接口
  • 如果 regex 长度大于 1 或包含元字符Stringsplit() 方法必须先将 regex 编译为 Pattern 对象,再调用 Patternsplit() 方法。这时,为避免重复编译,应该优先采用 Patternsplit() 方法
  • 如果 regex 就是一个字符且不是元字符Stringsplit() 方法会采用更为简单高效的实现。所以,这时应该优先采用 Stringsplit() 方法

5. 怎样理解 Stringmatches() 方法 public boolean matches(String regex)

  • Stringmatches() 方法实际调用的是 Patternmatches() 方法

    1
    2
    3
    4
    5
    public static boolean matches(String regex, CharSequence input) {
    Pattern p = Pattern.compile(regex); //编译 regex 为 Pattern 对象
    Matcher m = p.matcher(input); //调用 matcher() 方法生成一个匹配对象 Matcher
    return m.matches(); //Matcher 的 matches() 方法返回是否完整匹配
    }

6. 怎样理解 Matcher 的查找方法 find()

  • Matcher内部纪录有一个位置,起始为 0,find() 方法从这个位置查找匹配正则表达式的子字符串。找到后,返回 true,并更新这个内部位置,匹配到的子字符串信息可以通过如下方法获取

    • public String group():匹配到的完整子字符串
    • public int start():子字符串在整个字符串中的起始位置
    • public int end():子字符串在整个字符串中的结束位置加 1
  • group() 方法其实调用的是 group(0) 方法,表示获取匹配的第 0 个分组的内容。分组 0 是一个特殊的捕获分组,表示匹配的整个子字符串。除了分组 0,Matcher 还有如下方法,获取分组的更多信息

    • public int groupCount():分组个数
    • public String group(int group):分组编号为 group 的内容
    • public String group(String name):分组命名为 name 的内容
    • public int start(int group):分组编号为 group 的起始位置
    • public int end(int group):分组编号为 group 的结束位置加 1

7. 怎样理解 String 的替换方法

  • String 有多个替换方法

    • public String replace(char oldChar, char newChar)
    • public String replace(CharSequence target, CharSequence replacement)
    • pubilc String replaceAll(String regex, String replacement)
    • public String replaceFirst(String regex, String replacement)
  • 第一个 replace() 方法操作的是单个字符,第二个是 CharSequence,它们都是将参数看作普通字符。而 replaceAll()replaceFirst() 则将参数 regex 看作正则表达式,它们的区别是,replaceAll() 替换所有找到的子字符串,而 replaceFirst() 则只替换第一个找到的

    • replaceAll()replaceFirst() 中,参数 replacement 也不是被看作普通的字符串,可以使用美元符号加数字的形式(比如 $1)引用捕获分组。字符 $replacement 中是元字符,如果需要替换为字符 $ 本身,需要使用转义
    • 如果替换字符串是用户提供的,为避免元字符的干扰,可以使用 Matcher 的如下静态方法将其视为普通字符串:public static String quoteReplacement(String s)
  • StringreplaceAll()replaceFirst() 调用的其实是 PatternMatcher 中的方法。比如,replaceAll() 方法的代码是:public String replaceAll(String regex, String replacement) { return Pattern.compile(regex).matcher(this).replaceAll(replacement); }

  • replaceAll()replaceFirst() 都定义在 Matcher 中。除了一次性的替换操作外,Matcher 还定义了边查找、边替换的方法public Matcher appendReplacement(StringBuffer sb, String replacement)public StringBuilder appendTail(StringBuffer sb)。这两个方法用于和 find() 方法一起用

8. 写出下面代码的输出结果并分析

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
public static void replaceCat() {
Pattern p = Pattern.compile("cat");
Matcher m = p.matcher("one cat, two cat, three cat");
StringBuffer sb = new StringBuffer(); //存放最终的替换结果
int foundNum = 0;
while(m.find()) {
//Matcher 内部除了有一个查找位置,还有一个 append 位置,初始为 0,当找到一个匹配的子字符串后,appendReplacement() 做了三件事:
//1) 将 append 位置到当前匹配之前的子字符串 append 到 sb 中,在第一次操作中为 "one",第二次为 ", two"
//2) 将替换字符串 append 到 sb 中
//3) 更新 append 位置为当前匹配之后的位置
m.appendReplacement(sb, "dog");
foundNum++;
if(foundNum == 2) {
break;
}
}
m.appendTail(sb); //将 append 之后所有的字符 append 到 sb 中
System.out.println(sb.toString());
}

//输出:one dog, tow dog, three dog
//分析见注释
-------------------- 本文结束感谢您的阅读 --------------------