[原创]Unix下截取字符串_VMware, Unix及操作系统讨论区_Weblogic技术|Tuxedo技术|中间件技术|Oracle论坛|JAVA论坛|Linux/Unix技术|hadoop论坛

总帖数

每页帖数

1/1页

返回列表

发起投票

查看: 6118 | 回复: 0

主题： [原创]Unix下截取字符串

wulcan

版主

等级：中校
经验：1795
发帖：124
精华：0
注册：2014-3-19
状态：离线
发送短消息息给wulcan

加好友发送短消息息给wulcan

发消息

发表于：

2014-5-5 21:21:37 | [全部帖] [楼主帖]

楼主

截取字符串本身是一个很大的话题，在各shell操作中也经常遇到，高手们最常用的是grep、sed、awk组合，这三个命令都非常复杂，每一个都可以展开为一本书。
但平时的操作还真用很少用到复杂的表达式，追求简单是我们的不懈努力^_^
来看几个简单的截取：
１、截取最后部分的字符：
先看最后10行，这在看log的时候最常用：

tail /var/log/message
Apr 21 15:17:59 mm kernel: igb3: link state changed to DOWN
Apr 21 15:17:59 mm kernel: igb3: link state changed to UP
Apr 26 03:27:10 mm kernel: igb3: link state changed to DOWN
Apr 26 03:27:10 mm kernel: igb1: link state changed to DOWN
Apr 26 08:08:44 mm kernel: igb3: link state changed to UP
Apr 26 08:08:47 mm kernel: igb3: link state changed to DOWN
Apr 26 08:12:31 mm kernel: igb3: link state changed to UP
Apr 26 08:13:02 mm kernel: igb3: link state changed to DOWN
Apr 26 08:13:31 mm kernel: igb3: link state changed to UP
Apr 26 08:15:53 mm kernel: igb1: link state changed to UP

默认就是10行，要看50行呢，就加个-n参数吧：

tail –n 50 /var/log/message
Apr 10 07:12:29 mm kernel: igb3: link state changed to DOWN
……（48行）
Apr 26 08:15:53 mm kernel: igb1: link state changed to UP

这10行是按间顺序增大排列的，按时间递减可以不？

tail –rn 10 /var/log/message
Apr 26 08:15:53 mm kernel: igb1: link state changed to UP
Apr 26 08:13:31 mm kernel: igb3: link state changed to UP
Apr 26 08:13:02 mm kernel: igb3: link state changed to DOWN
Apr 26 08:12:31 mm kernel: igb3: link state changed to UP
Apr 26 08:08:47 mm kernel: igb3: link state changed to DOWN
Apr 26 08:08:44 mm kernel: igb3: link state changed to UP
Apr 26 03:27:10 mm kernel: igb1: link state changed to DOWN
Apr 26 03:27:10 mm kernel: igb3: link state changed to DOWN
Apr 24 22:02:42 mm kernel: ipfw: limit 80000 reached on entry 40
Apr 21 15:17:59 mm kernel: igb3: link state changed to UP

上面是以行计的，想看最后100个字符：

tail –c 100 /var/log/message
mm kernel: igb3: link state changed to UP
Apr 26 08:15:53 mm kernel: igb1: link state changed to UP

字符倒着输出应该更好玩吧，试试看：

Apr 26 08:15:53 mm kernel: igb1: link state changed to UP
mm kernel: igb3: link state changed to UP

不好意思，行可以倒，但字符不能倒，否则love成了evol，你还能看懂吗？
2、截取前开头的字符
用head命令，语法跟tail相似，就不再重复了。
3、截取某一列
先看一段测试vlan用的arp列表，注意最前面的问号（？）后面有一个空格：

arp –na
? (192.168.196.36) at 40:61:86:84:47:2a on vlan536 permanent [vlan]
? (192.168.196.35) at 40:61:86:84:47:2a on vlan535 permanent [vlan]
? (192.168.196.34) at 40:61:86:84:47:2a on vlan534 permanent [vlan]
? (192.168.196.32) at 40:61:86:84:47:2a on vlan532 permanent [vlan]
? (192.168.196.29) at 40:61:86:84:47:2a on vlan529 permanent [vlan]
? (192.168.196.28) at 40:61:86:84:47:2a on vlan528 permanent [vlan]
? (192.168.196.27) at 40:61:86:84:47:2a on vlan527 permanent [vlan]
? (192.168.196.26) at 40:61:86:84:47:2a on vlan526 permanent [vlan]
? (192.168.196.25) at 40:61:86:84:47:2a on vlan525 permanent [vlan]
? (192.168.196.24) at 40:61:86:84:47:2a on vlan524 permanent [vlan]
? (192.168.196.23) at 40:61:86:84:47:2a on vlan523 permanent [vlan]
? (192.168.196.22) at 40:61:86:84:47:2a on vlan522 permanent [vlan]
? (192.168.196.21) at 40:61:86:84:47:2a on vlan521 permanent [vlan]
? (192.168.196.20) at 40:61:86:84:47:2a on vlan520 permanent [vlan]
? (192.168.196.19) at 40:61:86:84:47:2a on vlan519 permanent [vlan]
? (192.168.196.18) at 40:61:86:84:47:2a on vlan518 permanent [vlan]
? (192.168.196.17) at 40:61:86:84:47:2a on vlan517 permanent [vlan]
? (192.168.196.16) at 40:61:86:84:47:2a on vlan516 permanent [vlan]

先看简单的例子：取出第3（注意不是前3）个字符：

arp –na | cut –b 3
(

(

(

(

……

显示一串左小括号。怎么回事呢？
首先看-b参数，cut支持三种模式：

-b：字节模式，按字节取出
-c：字符模式，按字符取出
-f：字段模式，按字段取出

字节和字符有什么区别呢？这主要用在多字节语言上，比如中文的GB2312编码，两个字节一个字符，-c将以两个字节为单位取出字符。举个例子，首先看当前的编码：

tt@v:~ % locale

LANG=

LC_CTYPE="zh_CN.GB2312"

LC_COLLATE="zh_CN.GB2312"

LC_TIME="zh_CN.GB2312"

LC_NUMERIC="zh_CN.GB2312"

LC_MONETARY="zh_CN.GB2312"

LC_MESSAGES="zh_CN.GB2312"

LC_ALL=zh_CN.GB2312

准备了一个中文文件：

tt@v:~ % cat test.txt

中华人民共和国山东省

来看-b模式，特殊字符，打印不出来，只好抓图显示了：

北京联动北方科技有限公司

会发现这是一个特殊字符，实际上，这是“中”字的后一个字节。再用-c看看：

tt@v:~ % cat test.txt | cut -c 2

华

这回出现“华”字了。
要截取一段字��怎么办呢？比如想把“共和国”三个字取出来？那就加上范围吧：

hu@vpn:~ % cat test.txt | cut -c 5-7

共和国

要再加上一个“中”字呢？，那就用逗号分成列表吧：

hu@vpn:~ % cat test.txt | cut -c 1,5-7

中共和国

再回到我们最初的例子：我想取出mac列（按数据库的概念来说，也叫字段），怎么办呢？要用到字段模式（-f）了：

arp –na | cut –f 4

? (192.168.196.36) at 40:61:86:84:47:2a on vlan536 permanent [vlan]
? (192.168.196.35) at 40:61:86:84:47:2a on vlan535 permanent [vlan]
? (192.168.196.34) at 40:61:86:84:47:2a on vlan534 permanent [vlan]
? (192.168.196.32) at 40:61:86:84:47:2a on vlan532 permanent [vlan]
? (192.168.196.29) at 40:61:86:84:47:2a on vlan529 permanent [vlan]
? (192.168.196.28) at 40:61:86:84:47:2a on vlan528 permanent [vlan]
? (192.168.196.27) at 40:61:86:84:47:2a on vlan527 permanent [vlan]
? (192.168.196.26) at 40:61:86:84:47:2a on vlan526 permanent [vlan]
? (192.168.196.25) at 40:61:86:84:47:2a on vlan525 permanent [vlan]
? (192.168.196.24) at 40:61:86:84:47:2a on vlan524 permanent [vlan]
? (192.168.196.23) at 40:61:86:84:47:2a on vlan523 permanent [vlan]
? (192.168.196.22) at 40:61:86:84:47:2a on vlan522 permanent [vlan]
? (192.168.196.21) at 40:61:86:84:47:2a on vlan521 permanent [vlan]
? (192.168.196.20) at 40:61:86:84:47:2a on vlan520 permanent [vlan]
? (192.168.196.19) at 40:61:86:84:47:2a on vlan519 permanent [vlan]
? (192.168.196.18) at 40:61:86:84:47:2a on vlan518 permanent [vlan]
? (192.168.196.17) at 40:61:86:84:47:2a on vlan517 permanent [vlan]
? (192.168.196.16) at 40:61:86:84:47:2a on vlan516 permanent [vlan]

咦？好像没有效果啊！
来读一下参数：
-f 是字段模式，这个没��，4是第四列，注意到问号单独占了一列，cut的序号又是从1开始的，所以也没错？哪地方出问题了呢？
问题出一“分隔符”上，用-f参数，必须指定分隔符，表示用哪个字符分隔字段，默认是制表字符（TAB），arp显示的结果，是用空格分隔的，所以要用-d参数指定空格作为分隔符。
但问题又来了，空格在命令行中不好用，会被shell解释为分隔符，cut –f 4 –d ，就跟丢了-d参数一样。
那就按shell的惯例，加上双引号吧，最终的命令是这样：

arp -na | cut -f 4 -d " "
40:61:86:84:47:2a
40:61:86:84:47:2a
40:61:86:84:47:2a
40:61:86:84:47:2a
40:61:86:84:47:2a
40:61:86:84:47:2a
……

这次，arp出来了。
怎么样，这三个命令还是比较好用的吧。

该贴由hui.chen转至本版2014-11-5 17:04:59