1. Sed简介                                                             
   sed 是一种在线编辑器，它一次处理一行内容。处理时，把当前处理的行存储
在临时缓冲区中，称为“模式空间”（pattern space），接着用sed命令处理缓
冲区中的内容，处理完成后，把缓冲区的内容送往屏幕。接着处理下一行，这样
不断重复，直到文件末尾。文件内容并没有 改变，除非你使用重定向存储输出。
Sed主要用来自动编辑一个或多个文件；简化对文件的反复操作；编写转换程序等
。以下介绍的是Gnu版本的Sed 3.02。                                      
                                             
2. 定址                                                              
可以通过定址来定位你所希望编辑的行，该地址用数字构成，用逗号分隔的两个
行数表示以这两行为起止的行的范围（包括行数表示的那两行）。如1，3表示1，
2，3行，美元符号($)表示最后一行。范围可以通过数据，正则表达式或者二者结
合的方式确定 。                                                        
                                                                
3. Sed命令                                                  
调用sed命令有两种形式：                                               
*                                                                      
sed [options] 'command' file(s)                                       
*   
sed [options] -f scriptfile file(s) 
a\                                                                     
在当前行后面加入一行文本。                                             
b lable                                                                
分支到脚本中带有标记的地方，如果分支不存在则分支到脚本的末尾。         
c\                                                                     
用新的文本改变本行的文本。                                     
d                                                                      
从模板块（Pattern space）位置删除行。                                  
D                                                            
删除模板块的第一行。                                                 
i\                                                                   
在当前行上面插入文本。                                                
h                                                  
拷贝模板块的内容到内存中的缓冲区。 
H                 
追加模板块的内容到内存中的缓冲区                                       
g                                                                      
获得内存缓冲区的内容，并替代当前模板块中的文本。                       
G                                                                      
获得内存缓冲区的内容，并追加到当前模板块文本的后面。                   
l                                                                      
列表不能打印字符的清单。                                               
n                                                                      
读取下一个输入行，用下一个命令处理新的行而不是用第一个命令。           
N                                                                      
追加下一个输入行到模板块后面并在二者间嵌入一个新行，改变当前行号码。   
p                                                                      
打印模板块的行。                                                       
P（大写）                                                              
打印模板块的第一行。                                                   
q                                                                      
退出Sed。                                                              
r file                                                                 
从file中读行。                                                         
t label                                                                
if分支，从最后一行开始，条件一旦满足或者T，t命令，将导致分支到带有标号
的命令处，或者到脚本的末尾。                                           
T label                                                                
错误分支，从最后一行开始，一旦发生错误或者T，t命令，将导致分支到带有标
号的命令处，或者到脚本的末尾。                                         
w file                                                                 
写并追加模板块到file末尾。                                             
W file                                                                 
写并追加模板块的第一行到file末尾。                                     
!                                                                      
表示后面的命令对所有没有被选定的行发生作用。                           
s/re/string                                                            
用string替换正则表达式re。                                             
=                                                                      
打印当前行号码。                                                       
#                                                                      
把注释扩展到下一个换行符以前。                                         
以下的是替换标记                                                       
*                                                                      
g表示行内全面替换。                                                    
*                                                                      
p表示打印行。                                                          
*                                                                      
w表示把行写入一个文件。                                                
*                                                                      
x表示互换模板块中的文本和缓冲区中的文本。                              
*                                                                      
y表示把一个字符翻译为另外的字符（但是不用于正则表达式）                
                                                                       
                                                                       
4. 选项                                                                
-e command, --expression=command                                       
允许多台编辑。                                                         
-h, --help                                                             
打印帮助，并显示bug列表的地址。                                        
-n, --quiet, --silent                                                  
取消默认输出。                                                         
-f, --filer=script-file                                                
引导sed脚本文件名。                                                    
-V, --version                                                          
打印版本和版权信息。                                                   
5. 元字符集^                                                           
锚定行的开始 如：/^sed/匹配所有以sed开头的行。                         
$                                                                      
锚定行的结束 如：/sed$/匹配所有以sed结尾的行。                         
匹配一个非换行符的字符 如：/s.d/匹配s后接一个任意字符，然后是d。       
*                                                                      
匹配零或多个字符 如：/*sed/匹配所有模板是一个或多个空格后紧跟sed的行。 
[]                                                                     
匹配一个指定范围内的字符，如/[Ss]ed/匹配sed和Sed。                     
[^]                                                                    
匹配一个不在指定范围内的字符，如：/[^A-RT-Z]ed/匹配不包含A-R和T-Z的一个
字母开头，紧跟ed的行。                                                 
\(..\)                                                                 
保存匹配的字符，如s/\(love\)able/\1rs，loveable被替换成lovers。        
&                                                                      
保存搜索字符用来替换其他字符，如s/love/**&**/，love这成**love**。      
\<                                                                     
锚定单词的开始，如:/\<love/匹配包含以love开头的单词的行。              
\>                                                                     
锚定单词的结束，如/love\>/匹配包含以love结尾的单词的行。               
x\{m\}                                                                 
重复字符x，m次，如：/0\{5\}/匹配包含5个o的行。                         
x\{m,\}                                                                
重复字符x,至少m次，如：/o\{5,\}/匹配至少有5个o的行。                   
x\{m,n\}                                                               
重复字符x，至少m次，不多于n次，如：/o\{5,10\}/匹配5--10个o的行。       
                                                                       
                                                                       
6. 实例                                                                
删除：d命令                                                            
*                                                                      
$ sed '2d' example-----删除example文件的第二行。                       
*                                                                      
$ sed '2,$d' example-----删除example文件的第二行到末尾所有行。         
*                                                                      
$ sed '$d' example-----删除example文件的最后一行。                     
*                                                                      
$ sed '/test/'d example-----删除example文件所有包含test的行。          
替换：s命令                                                            
*                                                                      
$ sed 's/test/mytest/g' example-----在整行范围内把test替换为mytest。如
果没有g标记，则只有每行第一个匹配的test被替换成mytest。                
*                                                                      
$ sed -n 's/^test/mytest/p' example-----(-n)选项和p标志一起使用表示只打
印那些发生替换的行。也就是说，如果某一行开头的test被替换成mytest，就打
印它。                                                                 
*                                                                      
$ sed 's/^192.168.0.1/&localhost/' example-----&符号表示替换换字符串中
被找到的部份。所有以192.168.0.1开头的行都会被替换成它自已加 localhost，
变成192.168.0.1localhost。                                             
*                                                                      
$ sed -n 's/\(love\)able/\1rs/p' example-----love被标记为1，所有
loveable会被替换成lovers，而且替换的行会被打印出来。                   
*                                                                      
$ sed 's#10#100#g' example-----不论什么字符，紧跟着s命令的都被认为是新
的分隔符，所以，“#”在这里是分隔符，代替了默认的“/”分隔符。表示把所
有10替换成100。                                                        
                                                                       
选定行的范围：逗号                                                     
*                                                                      
$ sed -n '/test/,/check/p' example-----所有在模板test和check所确定的范
围内的行都被打印。                                                     
*                                                                      
$ sed -n '5,/^test/p' example-----打印从第五行开始到第一个包含以test开
始的行之间的所有行。                                                   
*                                                                      
$ sed '/test/,/check/s/$/sed test/' example-----对于模板test和west之间
的行，每行的末尾用字符串sed test替换。                                 
                                                                       
                                                                       
多点编辑：e命令                                                        
*                                                                      
$ sed -e '1,5d' -e 's/test/check/' example-----(-e)选项允许在同一行里执
行多条命令。如例子所示，第一条命令删除1至5行，第二条命令用check替换test
。命令的执 行顺序对结果有影响。如果两个命令都是替换命令，那么第一个替换
命令将影响第二个替换命令的结果。                                       
*                                                                      
$ sed --expression='s/test/check/' --expression='/love/d' example-----
一个比-e更好的命令是--expression。它能给sed表达式赋值。                
                                                                      
从文件读入：r命令                                                      
*                                                                      
$ sed '/test/r file' example-----file里的内容被读进来，显示在与test匹配
的行后面，如果匹配多行，则file的内容将显示在所有匹配行的下面。         
写入文件：w命令                                                        
*                                                                      
$ sed -n '/test/w file' example-----在example中所有包含test的行都被写入
file里。                                                               
                                                                       
追加命令：a命令                                                        
*                                                                      
$ sed '/^test/a\\--->this is a example' example<-----'this is a
example'被追加到以test开头的行后面，sed要求命令a后面有一个反斜杠。     
                                                                       
插入：i命令                                                            
$ sed '/test/i\\                                                       
new line                                                               
------------------------' example                                      
                                                                       
如果test被匹配，则把反斜杠后面的文本插入到匹配行的前面。               
                                                                       
下一个：n命令                                                          
*                                                                      
$ sed '/test/{ n; s/aa/bb/; }' example-----如果test被匹配，则移动到匹配
行的下一行，替换这一行的aa，变为bb，并打印该行，然后继续。             
                                                                       
变形：y命令                                                            
*                                                                      
$ sed '1,10y/abcde/ABCDE/' example-----把1--10行内所有abcde转变为大写，
注意，正则表达式元字符不能使用这个命令。                               
退出：q命令                                                            
*                                                                      
$ sed '10q' example-----打印完第10行后，退出sed。                      
                                                                       
保持和获取：h命令和G命令                                               
*                                                                      
$ sed -e '/test/h' -e '$G example-----在sed处理文件的时候，每一行都被保
存在一个叫模式空间的临时缓冲区中，除非行被删除或者输出被取消，否则所有
被处理的行都将 打印在屏幕上。接着模式空间被清空，并存入新的一行等待处理
。在这个例子里，匹配test的行被找到后，将存入模式空间，h命令将其复制并存
入一个称为保 持缓存区的特殊缓冲区内。第二条语句的意思是，当到达最后一行
后，G命令取出保持缓冲区的行，然后把它放回模式空间中，且追加到现在已经存
在于模式空间中 的行的末尾。在这个例子中就是追加到最后一行。简单来说，任
何包含test的行都被复制并追加到该文件的末尾。
保持和互换：h命令和x命令                                               
*                                                                      
$ sed -e '/test/h' -e '/check/x' example -----互换模式空间和保持缓冲区
的内容。也就是把包含test与check的行互换。                              
                                                                       
7. 脚本                                                                
Sed脚本是一个sed的命令清单，启动Sed时以-f选项引导脚本文件名。Sed对于脚
本中输入的命令非常挑剔，在命令的末尾不能有任何空白或文本，如果在一行中
有多个命令，要用分号分隔。以#开头的行为注释行，且不能跨行。