shell对文件的操作

• 简介
• sed命令
• sed小技巧

---

一、简介

在shell脚本编写中，时常会用到对文件的相关操作，比如增加内容，修改内容，删除部分内容，查看部分内容等，但是上述举例的这些操作一般都是需要在文本编辑器中才能操作，常用的文本编辑器如：gedit、vim、nano等又是交互式文本编辑器，脚本无法自己独立完成，必须有人参与才可以完成。如果这样的话又违背了我们编写脚本的意愿(全部由机器来完成，减少人的工作压力，提升工作效率)。emm....如何才能让这些操作全部脚本自己就搞定，而不需要人的参与，而且又能按照我们的脚本预案来完成呢？

为了解决上述问题，linux为大家提供了一些命令,比如Perl、sed等命令，今天我就着重为大家介绍一下sed命令。

二、sed命令

sed是linux中提供的一个外部命令,它是一个行(流)编辑器，非交互式的对文件内容进行增删改查的操作，使用者只能在命令行输入编辑命令、指定文件名，然后在屏幕上查看输出。它和文本编辑器有本质的区别。

区别是：

文本编辑器: 编辑对象是文件

行编辑器：编辑对象是文件中的行

也就是前者一次处理一个文本，而后者是一次处理一个文本中的一行。这个是我们应该弄清楚且必须牢记的，否者可能无法理解sed的运行原理和使用精髓。

2.1)sed数据处理原理 image 2.2)sed命令

sed 命令

语法：

sed [options] '{command}[flags]' [filename]

命令选项
-e script 将脚本中指定的命令添加到处理输入时执行的命令中  多条件，一行中要有多个操作
-f script 将文件中指定的命令添加到处理输入时执行的命令中
-n        抑制自动输出
-i        编辑文件内容
-i.bak    修改时同时创建.bak备份文件。
-r        使用扩展的正则表达式
!         取反 （跟在模式条件后与shell有所区别）

sed常用内部命令
a   在匹配后面添加
i   在匹配前面添加
p   打印
d   删除
s   查找替换
c   更改
y   转换   N D P 

flags
数字             表示新文本替换的模式
g：             表示用新文本替换现有文本的全部实例
p：             表示打印原始的内容
w filename:     将替换的结果写入文件

2.2.1）sed内部命令说明

演示实例文档

[root@www ~]# cat data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

文件内容增加操作,将数据追加到某个位置之后，使用命令a。

演示案例

在data1的每行后追加一行新数据内容: append data "haha"
[root@www ~]# sed 'a\append data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
append data "haha"
2 the quick brown fox jumps over the lazy dog.
append data "haha"
3 the quick brown fox jumps over the lazy dog.
append data "haha"
4 the quick brown fox jumps over the lazy dog.
append data "haha"
5 the quick brown fox jumps over the lazy dog.
append data "haha"

在第二行后新开一行追加数据: append data "haha"
[root@www ~]# sed '2a\append data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
append data "haha"
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

在第二到四行每行后新开一行追加数据: append data "haha"
[root@www ~]# sed '2,4a\append data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
append data "haha"
3 the quick brown fox jumps over the lazy dog.
append data "haha"
4 the quick brown fox jumps over the lazy dog.
append data "haha"
5 the quick brown fox jumps over the lazy dog.

匹配字符串追加: 找到包含"3 the"的行，在其后新开一行追加内容: append data "haha"
[root@www ~]# sed '/3 the/a\append data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
append data "haha"
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

//开启匹配模式  /要匹配的字符串/

文件内容增加操作,将数据插入到某个位置之前，使用命令i。

演示案例

在data1的每行前插入一行新数据内容: insert data "haha"
[root@www ~]# sed 'i\insert data "haha"' data1
insert data "haha"
1 the quick brown fox jumps over the lazy dog.
insert data "haha"
2 the quick brown fox jumps over the lazy dog.
insert data "haha"
3 the quick brown fox jumps over the lazy dog.
insert data "haha"
4 the quick brown fox jumps over the lazy dog.
insert data "haha"
5 the quick brown fox jumps over the lazy dog.

在第二行前新开一行插入数据: insert data "haha"
[root@www ~]# sed '2i\insert data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
insert data "haha"
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

在第二到四行每行前新开一行插入数据: insert data "haha"
[root@www ~]# sed '2,4i\insert data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
insert data "haha"
2 the quick brown fox jumps over the lazy dog.
insert data "haha"
3 the quick brown fox jumps over the lazy dog.
insert data "haha"
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

匹配字符串插入: 找到包含"3 the"的行，在其前新开一行插入内容: insert data "haha"
[root@www ~]# sed '/3 the/i\insert data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
insert data "haha"
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

文件内容修改操作--替换,将一行中匹配的内容替换为新的数据，使用命令s。

演示案例

从标准输出流中做替换，将test替换为text
[root@www ~]# echo "this is a test" |sed 's/test/text/'
this is a text

将data1中每行的dog替换为cat
[root@www ~]# sed 's/dog/cat/' data1
1 the quick brown fox jumps over the lazy cat.
2 the quick brown fox jumps over the lazy cat.
3 the quick brown fox jumps over the lazy cat.
4 the quick brown fox jumps over the lazy cat.
5 the quick brown fox jumps over the lazy cat.

将data1中第二行的dog替换为cat
[root@www ~]# sed '2s/dog/cat/' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy cat.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

将data1中第二到第四行的dog替换为cat
[root@www ~]# sed '2,4s/dog/cat/' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy cat.
3 the quick brown fox jumps over the lazy cat.
4 the quick brown fox jumps over the lazy cat.
5 the quick brown fox jumps over the lazy dog.

匹配字符串替换:将包含字符串"3 the"的行中的dog替换为cat
[root@www ~]# sed '/3 the/s/dog/cat/' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy cat.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

文件内容修改操作--更改,将一行中匹配的内容替换为新的数据，使用命令c。

演示案例

将data1文件中的所有行的内容更改为: change data "data"
[root@www ~]# sed 'c\change data "haha"' data1
change data "haha"
change data "haha"
change data "haha"
change data "haha"
change data "haha"

将data1文件第二行的内容更改为: change data "haha"
[root@www ~]# sed '2c\change data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
change data "haha"
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

将data1文件中的第二、三、四行的内容更改为：change data "haha"
[root@www ~]# sed '2,4c\change data "haha"' data1
1 the quick brown fox jumps over the lazy dog.
change data "haha"
5 the quick brown fox jumps over the lazy dog.

将data1文件中包含"3 the"的行内容更改为: change data "haha"
[root@www ~]# sed '/3 the/c\change data "data"' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
change data "data"
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

文件内容修改操作--字符转换，将一行中匹配的内容替换为新的数据，使用命令y。

演示案例

将data1中的a b c字符转换为对应的 A  B  C字符
[root@www ~]# sed 'y/abc/ABC/' data1
1 the quiCk Brown fox jumps over the lAzy dog.
2 the quiCk Brown fox jumps over the lAzy dog.
3 the quiCk Brown fox jumps over the lAzy dog.
4 the quiCk Brown fox jumps over the lAzy dog.
5 the quiCk Brown fox jumps over the lAzy dog.

文件内容删除，将文件中的指定数据删除，使用命令d。

演示案例

删除文件data1中的所有数据
[root@www ~]# sed 'd' data1

删除文件data1中的第三行数据
[root@www ~]# sed '3d' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

删除文件data1第三到第四行的数据
[root@www ~]# sed '3,4d' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

删除文件data1中包含字符串"3 the"的行
[root@www ~]# sed '/3 the/d' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

文件内容查看，将文件内容输出到屏幕，使用命令p。

演示案例

打印data1文件内容
[root@www ~]# sed 'p' data1
1 the quick brown fox jumps over the lazy dog.
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

打印data1文件第三行的内容
[root@www ~]# sed '3p' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

打印data1文件第二、三、四行内容
[root@www ~]# sed '2,4p' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

打印data1文件包含字符串"3 the"的行
[root@www ~]# sed '/3 the/p' data1
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog. 

可以看得出，打印内容是重复的行，原因是打印了指定文件内容一次，又将读入缓存的所有数据打印了一次，所以会看到这样的效果，
如果不想看到这样的结果，可以加命令选项-n抑制内存输出即可。

2.2.2）命令选项说明

在命令行中使用多个命令 -e

将brown替换为green dog替换为cat
[root@www ~]# sed -e 's/brown/green/;s/dog/cat/' data1
1 the quick green fox jumps over the lazy cat.
2 the quick green fox jumps over the lazy cat.
3 the quick green fox jumps over the lazy cat.
4 the quick green fox jumps over the lazy cat.
5 the quick green fox jumps over the lazy cat.

从文件读取编辑器命令 -f 适用于日常重复执行的场景

1）将命令写入文件
[root@www ~]# vim abc
s/brown/green/
s/dog/cat/
s/fox/elephant/

2）使用-f命令选项调用命令文件
[root@www ~]# sed -f abc data1 
1 the quick green elephant jumps over the lazy cat.
2 the quick green elephant jumps over the lazy cat.
3 the quick green elephant jumps over the lazy cat.
4 the quick green elephant jumps over the lazy cat.
5 the quick green elephant jumps over the lazy cat.

抑制内存输出 -n

打印data1文件的第二行到最后一行内容  $最后的意思
[root@www ~]# sed -n '2,$p' data1 
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

使用正则表达式 -r

打印data1中以字符串"3 the"开头的行内容
[root@www ~]# sed -n  -r '/^(3 the)/p' data1
3 the quick brown fox jumps over the lazy dog.

从上述的演示中，大家可以看出，数据处理只是在缓存中完成的，并没有实际修改文件内容，如果需要修改文件内容可以直接使用-i命令选项。在这里我需要说明的是-i是一个不可逆的操作，一旦修改，如果想复原就很困难，几乎不可能，所以建议大家在操作的时候可以备份一下源文件。-i命令选项提供了备份功能，比如参数使用-i.bak，那么在修改源文件的同时会先备份一个以.bak结尾的源文件，然后再进行修改操作。

1）查看文件列表，没有发现data1.bak
[root@www ~]# ls
abc  apache  data1  Dobby  file  node-v10.14.1  Python-3.7.1  soft1  vimset

2）执行替换命令并修改文件
[root@www ~]# sed -i.bak 's/brown/green/' data1

3）发现文件夹中多了一个data1.bak文件
[root@www ~]# ls
abc     data1      Dobby  node-v10.14.1  soft1
apache  data1.bak  file   Python-3.7.1   vimset

4）打印比较一下，发现data1已经被修改，data1.bak是源文件的备份。
[root@www ~]# cat data1
1 the quick green fox jumps over the lazy dog.
2 the quick green fox jumps over the lazy dog.
3 the quick green fox jumps over the lazy dog.
4 the quick green fox jumps over the lazy dog.
5 the quick green fox jumps over the lazy dog.
[root@www ~]# cat data1.bak 
1 the quick brown fox jumps over the lazy dog.
2 the quick brown fox jumps over the lazy dog.
3 the quick brown fox jumps over the lazy dog.
4 the quick brown fox jumps over the lazy dog.
5 the quick brown fox jumps over the lazy dog.

2.2.3）标志

演示文档
[root@www ~]# cat data2
1 the quick brown fox jumps over the lazy dog . dog
2 the quick brown fox jumps over the lazy dog . dog
3 the quick brown fox jumps over the lazy dog . dog
4 the quick brown fox jumps over the lazy dog . dog
5 the quick brown fox jumps over the lazy dog . dog

数字标志：此标志是一个非零正数，默认情况下，执行替换的时候，如果一行中有多个符合的字符串，如果没有标志位定义，那么只会替换第一个字符串，其他的就被忽略掉了，为了能精确替换，可以使用数字位做定义。

替换一行中的第二处dog为cat
[root@www ~]# sed 's/dog/cat/2' data2
1 the quick brown fox jumps over the lazy dog . cat
2 the quick brown fox jumps over the lazy dog . cat
3 the quick brown fox jumps over the lazy dog . cat
4 the quick brown fox jumps over the lazy dog . cat
5 the quick brown fox jumps over the lazy dog . cat

g标志:将一行中的所有符合的字符串全部执行替换

将data1文件中的所有dog替换为cat
[root@www ~]# sed 's/dog/cat/g' data2
1 the quick brown fox jumps over the lazy cat . cat
2 the quick brown fox jumps over the lazy cat . cat
3 the quick brown fox jumps over the lazy cat . cat
4 the quick brown fox jumps over the lazy cat . cat
5 the quick brown fox jumps over the lazy cat . cat

p标志：打印文本内容，类似于-p命令选项

[root@www ~]# sed  '3s/dog/cat/p' data2
1 the quick brown fox jumps over the lazy dog . dog
2 the quick brown fox jumps over the lazy dog . dog
3 the quick brown fox jumps over the lazy cat . dog
3 the quick brown fox jumps over the lazy cat . dog
4 the quick brown fox jumps over the lazy dog . dog
5 the quick brown fox jumps over the lazy dog . dog

w filename标志:将修改的内容存入filename文件中

[root@www ~]# sed  '3s/dog/cat/w text' data2
1 the quick brown fox jumps over the lazy dog . dog
2 the quick brown fox jumps over the lazy dog . dog
3 the quick brown fox jumps over the lazy cat . dog
4 the quick brown fox jumps over the lazy dog . dog
5 the quick brown fox jumps over the lazy dog . dog

可以看出，将修改的第三行内容存在了text文件中
[root@www ~]# cat text 
3 the quick brown fox jumps over the lazy cat . dog

三、sed小技巧

$= 统计文本有多少行

统计data2有多少行
[root@www ~]# sed -n '$=' data2
5

打印data2内容时加上行号
[root@www ~]# sed  '=' data2
1
1 the quick brown fox jumps over the lazy dog . dog
2
2 the quick brown fox jumps over the lazy dog . dog
3
3 the quick brown fox jumps over the lazy dog . dog
4
4 the quick brown fox jumps over the lazy dog . dog
5
5 the quick brown fox jumps over the lazy dog . dog

job：

案例一：DNS监测WEB服务状态，并根据其状态实现高可用解析，场景：通过DNS进行单域名多条A记录解析做负载均衡。

!/bin/bash

CP1=0
CP2=0
while :
do

tong

 ping -c1 192.168.18.240 > /dev/null
 if [ $? -eq 1 ] && [ $CP1 -eq 0 ]
    then
       sed -i '/192.168.18.240/s/^/;/' /var/named/baidu.zone
       /etc/init.d/named reload
       CP1=1
  fi

butong

 ping -c1 192.168.18.240 > /dev/null 
 if [ $? -eq 0 ] && [ $CP1 -eq 1 ]
    then
        sed -i '/192.168.18.240/s/;//' /var/named/baidu.zone
        /etc/init.d/named reload
        CP1=0
 fi 


ping -c1 192.168.18.241 > /dev/null
if [ $? -eq 1 ] && [ $CP2 -eq 0 ]
    then
      sed -i '/192.168.18.241/s/^/;/' /var/named/baidu.zone
      /etc/init.d/named reload
      CP2=1
fi
   ping -c1 192.168.18.241 > /dev/null
 if [ $? -eq 0 ] && [ $CP2 -eq 1 ]
    then
        sed -i '/192.168.18.241/s/;//' /var/named/baidu.zone
        /etc/init.d/named reload
        CP2=0
 fi

sleep 5
done