sed的详细用法demo

先创建一个文件test
内容为：

lijie1:one   lijie2:two  lijie3:threehehe1:one1   hehe2:two2 hehe3:three3

1.查找

[root@lijie hadoop]# sed -n /lijie/p testlijie1:one   lijie2:two  lijie3:three

2.对当前字段的删除

--匹配删除[root@lijie hadoop]# cat ./test | sed /lijie/dhehe1:one1   hehe2:two2 hehe3:three3--指定函数删除[root@lijie hadoop]# cat ./test | sed 2dlijie1:one   lijie2:two  lijie3:three--多行删除（1-4行）[root@lijie hadoop]# cat ./test | sed 1,4d--删除到最后[root@lijie hadoop]# cat ./test | sed "3,$"d

3.替换

--给替换第一个[root@lijie hadoop]# cat test | sed s/hehe/heihei/lijie1:one   lijie2:two  lijie3:threeheihei1:one1   hehe2:two2   hehe3:three3--替换所有[root@lijie hadoop]# cat test | sed s/hehe/heihei/glijie1:one   lijie2:two  lijie3:threeheihei1:one1   heihei2:two2 heihei3:three3

4.替换然后打印匹配行

[root@lijie hadoop]# cat test | sed -n /three3/s/hehe/heihei/gpheihei1:one1   heihei2:two2 heihei3:three3

5.复合条件-e（把lijie换成hadoop，把匹配three3的hehe换为heihei）

[root@lijie hadoop]# cat test | sed -n -e s/lijie/hadoop/gp -e /three3/s/hehe/heihei/gphadoop1:one   hadoop2:two  hadoop3:threeheihei1:one1   heihei2:two2 heihei3:three3

6.对文件进行修改

[root@lijie hadoop]# sed -i -n -e s/lijie/hadoop/gp -e /three3/s/hehe/heihei/gp test [root@lijie hadoop]# cat test hadoop1:one   hadoop2:two  hadoop3:threeheihei1:one1   heihei2:two2 heihei3:three3

7.将某个文件的文件替换到匹配的那一行后面(r和R不同)

--r[root@lijie hadoop]# sed /hadoop/'r test1' testhadoop1:one   hadoop2:two  hadoop3:threen1n2heihei1:one1   heihei2:two2 heihei3:three3--R[root@lijie hadoop]# sed /hadoop/'r test1' testhadoop1:one   hadoop2:two  hadoop3:threen1heihei1:one1   heihei2:two2 heihei3:three3hadoop4:one   hadoop5:two  hadoop6:threen2hadoop7:one   hadoop8:two  hadoop9:three

8.追加字符a为匹配行后，i为匹配行前

[root@lijie hadoop]# sed /hadoop/'a this is append str' testhadoop1:one   hadoop2:two  hadoop3:threethis is append strheihei1:one1   heihei2:two2 heihei3:three3[root@lijie hadoop]# sed /hadoop/'i this is append str' testthis is append strhadoop1:one   hadoop2:two  hadoop3:threeheihei1:one1   heihei2:two2 heihei3:three3

9.整行替换

[root@lijie hadoop]# sed /hadoop/'c new line' testnew lineheihei1:one1   heihei2:two2 heihei3:three3