A sample file demo.txt contains the following content

1
2
3
4
5
6
7
8
9
10
1 Linux Operating System
2 Unix Operating System
3 RHEL
4 Red Hat
5 Fedora
6 Arch Linux
7 CentOS
8 Debian
9 Ubuntu
10 openSUSE

How do we delete lines that contain operating?

1. Find the line number

1
2
3
$ grep -in "operating" demo.txt
1:1 Linux Operating System
2:2 Unix Operating System

2. Then use awk to get the first chunk

1
2
3
$ grep -in "operating" demo.txt | awk '{print $1}'
1:1
2:2

3. Get only the line number

This will split the string by delimiter :, and get the first chunk

1
2
3
$ grep -in "operating" demo.txt | awk '{print $1}' | cut -d ':' -f 1
1
2

4. Delete by using sed command

The sed command delete by line number like this

1
$ sed '23d;45d;102d' filename.ext

Before that we already can get line number, so now we have to join it into a single line

1
2
$ grep -in "operating" demo.txt | awk '{print $1}' | cut -d ':' -f 1 | xargs -I'{}' echo '{}d' | paste -d';' -s
1d;2d
  • with xargs, it append a character “d”
  • with paste, 2 lines joined into a single line

At the end, we can

1
2
3
4
5
6
7
8
9
$ grep -in "operating" demo.txt | awk '{print $1}' | cut -d ':' -f 1 | xargs -I'{}' echo '{}d' | paste -d';' -s | xargs -I'{}' sed '{}' demo.txt
3 RHEL
4 Red Hat
5 Fedora
6 Arch Linux
7 CentOS
8 Debian
9 Ubuntu
10 openSUSE

It will output to stdout, if you want to replace the file, use -i option

1
$ grep -in "operating" demo.txt | awk '{print $1}' | cut -d ':' -f 1 | xargs -I'{}' echo '{}d' | paste -d';' -s | xargs -I'{}' sed -i '{}' demo.txt

Make it into a shell script function

1
2
3
4
5
6
7
8
delete_lines_contain_text () {
    txt=$1
    file=$2
    grep -in $txt $file | awk '{print $1}' | cut -d ':' -f 1 | xargs -I'{}' echo '{}d' | paste -d';' -s | xargs -I'{}' sed -i '{}' $file
}

# usage
delete_lines_contain_text "operating" demo.txt

References: