正则表达式解惑

使用正则表达式(组的概念/拆分字符串)

正则表达式(Regular expressions)是一套语法匹配规则，各种语言，如Perl， .Net和Java都有其对应的共享的正则表达式类库。在.Net中，这个类库叫做Regex。

简单的说，Regex是从字符窗中查找匹配字符串的应用类。通过Regex，编程人员能够非常方便的从一段数据中提取自己所需要的数据信息。举一个简单的例子，让大家对Regex有个大概的了解：

 

结果很明显，regex为我们找到了字符串”fox 9212gold”中的数字字符串，输出结果为”9212” .

       对Regex有了一个基本的概念之后，我需要告诉你一个非常好的消息，那就是Regex可以为我们做的远不止这么一点点，他是一套功能非常强大语法匹配规则。当然，这里还有一个坏消息等着我们，那就是强大功能的语法规则自然需要大量纷繁复杂的keyword支持，这也为Regex的学习带来了极大的难度。想要真正掌握正则表达式，可不是几个sample能够全部揭示与说明的。

 

创建一个Regex对象

       Regex的构造函数有三种，在这里就不讨论默认构造函数了。另外两个构造函数中，一个构造函数接收正则表达式字符串作为入参，另一个以正则表达式字符串和RegexOptions作为入参。如：

Regex regex = new Regex("\w+$");

Regex regex = new Regex("\s+", RegexOptions.IgnoreCase | RegexOptions.Multiline);

RegexOptions可以为我们提供一些特殊的帮助，比如IgnoreCase能够在匹配是忽略大小写，Multiline能够调整^和$的意义，改为匹配一行的开头和结尾。

上面我们构造了一个正则表达式，只不过我们还没有用它来做任何事情，马上我们就可以通过使用下面的几个方法，实现对字符串对象的操作了。

匹配字符串

       Regex有两个获取匹配的方法Match()和Matches()，分别代表匹配一个，匹配多个结果。这里就用Matches来展示一下怎么使用Regex获取匹配字符串，并将其显示出来。

         public static void showMatches(string expression, RegexOptions option, string ms)                 

         {                                                                                                    

                   Regex regex = new Regex(expression, option);                                             

                   MatchCollection matches = regex.Matches(ms);                                           

                   //show matches                                                                                  

                   Console.WriteLine("////////////////----------------------------------////////////////");                     

                   Console.WriteLine(" string: \"{0}\"\r\n expression: \"{1}\"\r\n match result is:",             

                  ms, expression);                                                                                

                   foreach(Match m in matches)                                                          

                   {                                                                                                                        Console.WriteLine("match string is: \"{0}\", length: {1}",                       

                           m.Value.ToString(), m.Value.Length);                                            

                   }                                                                                           

                   Console.WriteLine("matched count: {0}", matches.Count);                                   

         }                                                                                                    

 

方法Matched通过比较入参字符串和正则表达式，找到所有符合的结果，并将结果作为MatchCollection传出来。这样，只要简单的遍历这个collection，就可以很快的获得所有的结果。

 

组的概念

 

       当你获得这样的一个字符串”最后比分是：19/24”，你肯定希望有一个正则表达式，他不单能够找到形如 data1/data2的字符串，还应该能够直接把data1,data2作为单独的结果传送出来。否则你需要再对形如”19/24”的字符串进行分析，才能够顺利得到双方的比分。显然，正则表达式不会忽略这个问题，所以他增加了组的概念。你可以把一次搜索的结果分别放入不同的组中，并通过组名或者组的所以分别取得这些组的结果。比如上例，我们就可以用@”(\d+)/(\d+)”作为表达式。来看看结果吧：

 

那么如何区分组呢？

用一对（）括起来的正则就是一个组。

例如：reg = ([a-z])(\.)(\d) 这个正则表达式有三对短号，就有三个组：[a-z] .和\d 这三个数

str = "a.4g"

匹配的结果为 a.4

对应的组为：

$0=a.4

$1=a

$2=.

$3=d

如果reg = ([a-z])\.\d 这个正则只有一个括号，就只有一个组[a-z]

str="a.4g"匹配结果为a.4,对应的组为

$0=a.4

$1=a

 

                            Regex regex = new Regex(@"(\d+)/(\d+)");                                             

                            MatchCollection matches = regex.Matches(@"最后比分是：19/24");        

                            //show matches                                                                         

                            Console.WriteLine("////////////////----------------------------------////////////////");            

                            foreach(Match m in matches)                                                 

                            {                                                                                  

                                     //Console.WriteLine("match string is: \"{0}\", length: {1}",                                                 // m.Value.ToString(), m.Value.Length);                                 

                                     foreach(string name in regex.GetGroupNames())                      

                                     {                                                                        

                                               Console.WriteLine("\r capture group \"{0}\" value is:\"{1}\""        

                                              , name, m.Groups[name].Value);                                 

                                     }                                                                        

                            }                                                                                  

                            Console.WriteLine("matched count: {0}", matches.Count);                         

 

输出：

////////////////----------------------------------////////////////

 capture group "0" value is:"19/24"

 capture group "1" value is:"19"

 capture group "2" value is:"24"

matched count: 1

 

现在清楚了吧，Regex对象把匹配的结果放入组0中。同时，匹配的组信息也放入了对应的组中。组的名称在默认的情况下，是从1开始依次增加的整数。0作为保留名称，专门用于指示匹配的整个字符串。既然有”默认情况”这样的概念，自然也就意味着用户可以自定义组的名称。方法很简单，在组的前面加上：?<name>就可以了。好了，现在把前面的正则表达式修改一下，换成@”(?<score1>\d+)/(?<score1>\d+)”，现在再看看结果：

////////////////----------------------------------////////////////

 capture group "0" value is:"19/24"

 capture group "score1" value is:"19"

 capture group "score2" value is:"24"

matched count: 1

 

换成自己定义的名字了吧，哈哈！为了在今后的测试中，能够更加方便的看到所有结果，我们对前面介绍过的showMatches()做一点小小的调整。这样，如果在表达式中包含了组的定义，我们也可以在不需要修改任何代码的情况下，直接看到所有的组信息了，调整后的方法showMatchesPro()如下：

         public static void showMatchesPro(string expression, RegexOptions option, string ms)          

         {                                                                                                    

                   Regex regex = new Regex(expression, option);                                             

                   MatchCollection matches = regex.Matches(ms);                                           

                   //show matches                                                                                  

                   Console.WriteLine("////////////////----------------------------------////////////////");                     

                   Console.WriteLine(" string: \"{0}\"\r\n expression: \"{1}\"\r\n match result is:",             

                  ms, expression);                                                                                

                   foreach(Match m in matches)                                                          

                   {                                                                                                                       foreach(string name in regex.GetGroupNames())                                

                            {                                                                                  

                                     Console.WriteLine("\r capture group \"{0}\" value is:\"{1}\"",                

                                     name, m.Groups[name].Value);                                            

                            }                                                                                  

                   }                                                                                           

                   Console.WriteLine("matched count: {0}", matches.Count);                                   

                   // show group name                                                                 

                  Console.WriteLine("group name count {0}", regex.GetGroupNames().Length);                 

                   foreach(string name in regex.GetGroupNames())                                         

                   {                                                                                           

                            Console.WriteLine("group name :\"{0}\"", name);                                 

                   }                                                                                           

         }                                                                                                    

 

替换字符串

Regex也提供了方便的匹配结果替换的功能。为了方便测试，我们也把他写成方法，代码如下：

        public static string replaceMatch(string expression, RegexOptions option, string ms, string rep)     

         {                                                                                                    

                   Regex regex = new Regex(expression, option);                                             

                   string result = regex.Replace(ms, rep);                                                  

                   Console.WriteLine("////////////////----------------------------------////////////////");                     

                   Console.WriteLine("string: \"{0}\", expression:\"{1}\", replace by : \"{2}\"",                    

                  ms, expression, rep);                                                                        

                   Console.WriteLine("replace result string is: \"{0}\", length: {1}",                                    

                                    result.ToString(), result.Length);                                            

                            return result;                                                                     

          }                                                                                                                                                                                                                

 

Regex.Replace通常接受两个string作为入参，第一个string为输入字符串。第二个字符串用来替代匹配字符串，它可以包含一些特殊字符串，用来代表特别转换。

 

特殊字符串                      替换结果

$&               匹配的字符串，也可以用$0

$1, $2, . . .    匹配字符串中的对应组，用索引标示

${name}      匹配字符串中的对应组，用名称标示

$‘              匹配位置之前的字符串

$’              匹配位置之后的字符串

$$         一个‘$’ 字符

$_          输入字符串

$+         匹配字符串的所有组中，最后一个组中的数据

 

是不是看了这么多怪型怪状的特殊字符串，有点头晕了？嗯，那就弄两个sample来看看结果吧！

Sample1:

replaceMatch(@"\d+", RegexOptions.None, "fef 12/21 df 33/14   727/1", "<<$&>>");

输出，所有数字型的数据都被替换成了<<data>>：

////////////////----------------------------------////////////////

string: "fef 12/21 df 33/14   727/1", expression:"\d+", replace by : "<<$&>>"

replace result string is: "fef <<12>>/<<21>> df <<33>>/<<14>>   <<727>>/<<1>>",

length: 50

 

Sample2:

replaceMatch(@"(\d+)/(\d+)", RegexOptions.None, "fef 12/21 df 33/14   727/1", "$+");

输出，所有data1/data2匹配的数据，都被替换成了data2：

////////////////----------------------------------////////////////

string: "fef 12/21 df 33/14   727/1", expression:"(\d+)/(\d+)", replace by : "$+"

replace result string is: "fef 21 df 14   1", length: 16

 

怎么样，Regex的功能够丰富的吧！可是，也许你的需求不光这么简单，比如说，你要把”I have 200 dollars”中间的money加倍，怎么办？我晕倒，好像没有现成的东西可以用。没有关系，Regex还有更好的功能。它允许你自己定义转换公式。

using System.Text.RegularExpressions;                                                                  

class RegularExpressions                                                                                          

{                                                                                                              

   static string CapText(Match m)                                                                              

   {                                                                                                        

         // Get the matched string.                                                                           

         string x = m.ToString();                                                                               

         // double this value                                                                             

         string result = (int.Parse(x) * 2).ToString();                                                                

         return result;                                                                                        

   }                                                                                                            

   static void Main()                                                                                            

   {                                                                                                        

         string text = "i have 200 dollars";                                                                        

         string result = Regex.Replace(text, @"\d+",                                                              

         new MatchEvaluator(RegularExpressions.CapText));                                                     

         System.Console.WriteLine("result=[" + result + "]");                                                

   }                                                                                                        

}                                                                                                              

 

看看结果，太好了，我的钱真的变成了两倍！

但本文的目的是希望提供给大家一个方便好用的测试类，因此我们重载上面的repalceMatch方法，也允许自定义转换公式作为入参：

         public static string replaceMatch(string expression, RegexOptions option, string ms,               

                                    MatchEvaluator evaluator)                                                       

                   {                                                                                           

                            Regex regex = new Regex(expression, option);                                    

                            string result = regex.Replace(ms, evaluator);                                        

                            Console.WriteLine("////////////////----------------------------------////////////////");            

                            Console.WriteLine("string: \"{0}\", expression:\"{1}\", replace by a evaluator.", ms,    

                                                       expression);                                          

                            Console.WriteLine("replace result string is: \"{0}\", length: {1}", result.ToString(),     

                                                       result.Length);                                      

                            return result;                                                                     

                   }                                                                                           

 

拆分字符串

 

Regex还提供了从匹配位置将字符串拆分的方法Split。这个方法同样具有多个重载，不过这些都不是重点，大家可以自己看文档。我们继续完成我们用于测试的方法：

                   public static void splitMatch(string expression, RegexOptions option, string ms)              

                   {                                                                                           

                            Regex regex = new Regex(expression, option);                                    

                            string[] result = regex.Split(ms);                                                      

                            Console.WriteLine("////////////////----------------------------------////////////////");            

                            Console.WriteLine("string: \"{0}\", expression: \"{1}\", split result is:", ms,                  

                                               expression);                                                   

                            foreach(string m in result)                                                                

                            {                                                                                  

                                     Console.WriteLine("splited string is: \"{0}\", length: {1}",           

                                              m.ToString(), m.Length);                                       

                            }                                                                                  

                            Console.WriteLine("splited count: {0}", result.Length);                                 

                   }                                                                                           

代码简单，不多做解释。直接来一个smaple看看结果：

splitMatch(@"/",RegexOptions.None,  "2004/4/25");

输出：

////////////////----------------------------------////////////////

string: "2004/4/25", expression: "/", split result is:

splited string is: "2004", length: 4

splited string is: "4", length: 1

splited string is: "25", length: 2

splited count: 3

 

结果很明显，收工。

 

我的目的</s

分享到：

Oracle 带条件插入/更新 | 修改oracle数据库字符集的方法

• 2010-08-24 13:47
• 浏览 628
• 评论(0)
• 分类:编程语言
• 查看更多