grep, sed, awk Notes

最新推荐文章于 2023-02-18 18:00:25 发布

EverNoob

最新推荐文章于 2023-02-18 18:00:25 发布

阅读量150

点赞数

分类专栏： Notes Linux

本文链接：https://blog.youkuaiyun.com/maxzcl/article/details/117470939

版权

Notes 同时被 2 个专栏收录

142 篇文章

订阅专栏

Linux

15 篇文章

订阅专栏

博客主要介绍了grep、sed、awk三个工具。grep可通过文本、通配符或正则表达式匹配；sed将文件读入流并逐行处理，可进行内容选择和编辑；awk是一种类似C的小型编程语言，能剖析文件并过滤特定模式或信息，还给出了相关详细解释和示例的链接。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

overview: https://sunjian.blog.youkuaiyun.com/article/details/82930982

grep

detailed explanation and examples: https://xie1997.blog.youkuaiyun.com/article/details/82974781

grep pattern ==> match by text or linux wildcards

grep "pattern" ==> match by regular expression with soft escape (variable, command replacement allowed)

grep 'pattern' ==> match by regular expression with hard escape (use as simple text)

for linux wildcards vs. linux (E)RE: https://blog.youkuaiyun.com/youmatterhsp/article/details/80528761

for generic ERE rules: https://blog.youkuaiyun.com/scgaliguodong123_/article/details/45363409

sed

for detailed explanation and examples: https://www.cnblogs.com/xuxiuxiu/p/6945385.html

1. sed read the file into a stream and process the file content line by line

2. sed option edit_command filename ==> "sed -p filename" will simply print the file line by line

===> all fancier manipulation of content selection and editting can be built up from here.

===> with neither output redirection nor -i option, NO sed command will edit the original file content; it will simply print the result to console.

3. (from https://sunjian.blog.youkuaiyun.com/article/details/82930982)

both

sed '#a txt' filename

sed '#a\txt' filename

work

4. 2 method for search/match and partially replace within a line:

from https://www.cnblogs.com/xuxiuxiu/p/6945385.html

One: Pipeline

Two:

change the 's' edit command to 'd' to delete lines with matched pattern.

==> unfortunately back reference cannot be parsed across '/'

e.g.

sed -i '/\(\(la\)\{2,\}\)/s/\1/hihihi/g' file // match lalala, replace lalala with hihihi

gives: invalid back reference

\(\) will group together expressions; not () (not for RE, though () should do the job for ERE. while \( matches the char '(' ).

la\{2,\} ==> laa.....

\(la\)\{2,\} ==> lala.......

awk

official doc for gawk: http://www.gnu.org/software/gawk/manual/gawk.html

intro and examples: https://www.cnblogs.com/ggjucheng/archive/2013/01/13/2858470.html

a mini programming language, a tool to disect the file and filter for particular patterns/info;

C-like; the most powerful of the 3 and can be cumbersome to deploy ==> use scripts to deal with complex instructions.