Mastering sed: Essential Commands for Fast Text Processing

sed is a widely used, simple yet powerful command‑line tool for fast text processing. It is considered a hard skill that every programmer should master because of its versatility and similarity to vim and regular‑expression syntax.

A Simple Introduction

A basic sed command consists of three parts: options, address range, and the editing command, followed by the target file. The

-f

option can load a script file (advanced usage, not covered here).

Parameters

-n

(or

--quiet

/

--silent

) suppresses the default output, printing only the results you request.

-i

edits files in place, which is very dangerous because it overwrites the original file.

Using -i will modify the original file directly; be extremely careful.

Address Ranges

Line numbers can specify ranges, e.g.,

1,4

selects lines 1‑4. Other examples:

<code>sed -n '5p' file</code><code>sed -n '2,5p' file</code><code>sed -n '1~2p' file</code><code>sed -n '2~2p' file</code><code>sed -n '2,+3p' file</code><code>sed -n '2,$p' file</code>

Regular‑expression ranges are also supported:

<code>sed -n '/sys/,+3p' file</code><code>sed -n '/^sys/,/mem/p' file</code>

Editing Commands

The most common command is

p

(print). Others include:

p prints matching lines. d deletes matching lines (requires dropping -n ). w writes matching lines to another file.

Less frequently used commands such as

a

,

i

, and

c

are omitted for brevity.

<code>sed -n '2,5p' file</code><code>sed '2,5d' file</code><code>sed -n '2,5w output.txt' file</code>

Substitution Mode

The

s

command replaces matched text. Its basic form is

s/pattern/replacement/flags

. Flags include

g

(global replace),

p

(print when used with

-n

),

w

(write changed lines),

i

(ignore case), and

e

(execute the result as a command).

<code>sed -n 's/a/b/gipw output.txt' file</code><code>sed 's/^/ls -la/e' file</code>

The special placeholder

&

represents the entire matched text, allowing constructions such as:

<code>sed 's/.*/"&amp;"/' file</code>

Alternative Delimiters

Because slashes often appear in patterns, you can use other delimiters:

<code>sed '/aaa/s@/etc@/usr@g' file</code><code>sed '/aaa/s|/etc|/usr|g' file</code>

In‑Place Editing with Backup

Use

-i.bak

to edit a file while keeping a backup:

<code>sed -i.bak 's/a/b/' file</code>

Practical One‑Liners

Print lines longer than 50 characters:

<code>sed -n '/^.{50}/p'</code>

Count word occurrences:

<code>sed 's/ /\n/g' file | sort | uniq -c</code>

Remove comments from Python files:

<code>find ./ -name "*.py" | xargs sed -i.bak '/^[ ]*#/d'</code>

Show specific line ranges:

<code>sed -n -e '5,7p' -e '10,13p' file</code>

Extract only IP addresses:

<code>ip route show | sed -n '/src/p' | sed -e 's/  */ /g' | cut -d' ' -f9</code>

sed also offers an

x

(exchange) command and can even be used to implement games like Sokoban ( GitHub link ).