正则表达式 java.util.regex Pattern

谁践踏了优雅 2022-09-23 14:49 169阅读 0赞

最近写字符串处理程序用到了Java正则表达式。其中写了个  \\\\s\*:\\\\s\*|(?<!Egress)\\\\s\{2,\}  。其语义是， “任意个空格+：+任意个空格”的形式或者 “两个以上的空格，但是前边不能是Egress”。

为了写正则表达式，特意查看了JDK帮助文档，其中有一段：

<table style="color:rgb(51,51,51); font-family:Arial; font-size:14px; line-height:26px"> 
 <tbody> 
 <tr> 
 <th colspan="2">Special constructs (non-capturing)</th> 
 </tr> 
 <tr> 
 <td>(?:X)</td> 
 <td>X, as a non-capturing group</td> 
 </tr> 
 <tr> 
 <td>(?idmsux-idmsux)&nbsp;</td> 
 <td>Nothing, but turns match flags&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#CASE_INSENSITIVE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">i</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#UNIX_LINES" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">d</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#MULTILINE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">m</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#DOTALL" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">s</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#UNICODE_CASE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">u</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#COMMENTS" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">x</a>&nbsp;on - off</td> 
 </tr> 
 <tr> 
 <td>(?idmsux-idmsux:X)&nbsp;&nbsp;</td> 
 <td>X, as a&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#cg" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">non-capturing group</a>&nbsp;with the given flags&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#CASE_INSENSITIVE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">i</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#UNIX_LINES" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">d</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#MULTILINE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">m</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#DOTALL" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">s</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#UNICODE_CASE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">u</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/j2se_1.6/api/java/util/regex/Pattern.html#COMMENTS" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">x</a>&nbsp;on - off</td> 
 </tr> 
 <tr> 
 <td>(?=X)</td> 
 <td>X, via zero-width positive lookahead</td> 
 </tr> 
 <tr> 
 <td>(?!X)</td> 
 <td>X, via zero-width negative lookahead</td> 
 </tr> 
 <tr> 
 <td>(?&lt;=X)</td> 
 <td>X, via zero-width positive lookbehind</td> 
 </tr> 
 <tr> 
 <td>(?&lt;!X)</td> 
 <td>X, via zero-width negative lookbehind</td> 
 </tr> 
 <tr> 
 <td>(?&gt;X)</td> 
 <td>X, as an independent, non-capturing group</td> 
 </tr> 
 </tbody> 
</table>

<table style="color:rgb(51,51,51); font-family:Arial; font-size:14px; line-height:26px"> 
 <tbody> 
 <tr> 
 <th colspan="2">特殊构造（非捕获）</th> 
 </tr> 
 <tr> 
 <td>(?:X)</td> 
 <td>X，作为非捕获组</td> 
 </tr> 
 <tr> 
 <td>(?idmsux-idmsux)&nbsp;</td> 
 <td>Nothing，但是将匹配标志<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#CASE_INSENSITIVE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">i</a><a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#UNIX_LINES" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">d</a><a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#MULTILINE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">m</a><a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#DOTALL" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">s</a><a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#UNICODE_CASE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">u</a><a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#COMMENTS" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">x</a>&nbsp;on - off</td> 
 </tr> 
 <tr> 
 <td>(?idmsux-idmsux:X)&nbsp;&nbsp;</td> 
 <td>X，作为带有给定标志&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#CASE_INSENSITIVE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">i</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#UNIX_LINES" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">d</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#MULTILINE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">m</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#DOTALL" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">s</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#UNICODE_CASE" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">u</a>&nbsp;<a href="http://static.oschina.net/uploads/doc/javase-6-doc-api-zh_CN/java/util/regex/Pattern.html#COMMENTS" style="color:rgb(51,102,153); text-decoration:none" rel="nofollow">x</a>&nbsp;on - off</td> 
 </tr> 
 <tr> 
 <td>(?=X)</td> 
 <td>X，通过零宽度的正 lookahead</td> 
 </tr> 
 <tr> 
 <td>(?!X)</td> 
 <td>X，通过零宽度的负 lookahead</td> 
 </tr> 
 <tr> 
 <td>(?&lt;=X)</td> 
 <td>X，通过零宽度的正 lookbehind</td> 
 </tr> 
 <tr> 
 <td>(?&lt;!X)</td> 
 <td>X，通过零宽度的负 lookbehind</td> 
 </tr> 
 <tr> 
 <td>(?&gt;X)</td> 
 <td>X，作为独立的非捕获组</td> 
 </tr> 
 </tbody> 
</table>

好家伙，这文档写的，英文不是中文，中文也不是中文。不得不解释下鸟：

lookahead和lookbehind也就是先行和后行。

先行指的是，在进行断言的时候，字符串匹配引擎指针会被向前移动以进行断言（由于zero-width，所以之后会再被移回）。

后行指的是，在进行断言的时候，字符串匹配引擎指针会被向后移动以进行断言（由于zero-width，所以之后会再被移回）。

zero-width也就是零宽度。

指的是断言前后不会改变当前字符串匹配引擎的指针位置。

positive 和negative  表示是预测匹配还是不匹配。

因此：

1 (?=X)，预测其前边字串匹配位置后为X。

(?=*X*)的意思就是预测匹配位置之后出现的应该是X。如果符合则匹配上。在(?=X)后边带上不以X为开始的任意字符串都将导致该表达式不能被匹配。因为预测不改变当前字符串搜索的指针（零宽度）。

例如：

表达式"a(?=b)c"不能匹配"abc"，也不能匹配"ac"、"ab"。而且不能匹配任何字符串。因为其预测匹配a的位置后，应该是b，但是，又要求a之后是c。