Mastering sed: Essential Commands, Options, and Real‑World Examples

What is sed?

sed (stream editor) reads input line by line, stores each line in a temporary buffer called the pattern space , applies editing commands to lines that match a given address, and prints the result. By default the original file is not modified; in‑place editing requires the -i option.

Basic syntax

sed [OPTION]... 'script' [input-file]...

Common options

-n

, --quiet, --silent: suppress automatic printing of the pattern space (useful with p). --debug: annotate program execution for debugging. -e script, --expression=script: specify a script on the command line; multiple -e arguments are allowed. -f script-file, --file=script-file: read commands from a file. -i[SUFFIX], --in-place[=SUFFIX]: edit files in place; if SUFFIX is given a backup copy is created. -l N, --line-length=N: set line‑wrap length for the l command. --posix: disable GNU extensions and follow POSIX syntax. -E, -r, --regexp-extended: enable extended regular expressions (ERE). -s, --separate: treat each input file as a separate stream. --sandbox: run in sandbox mode, disabling e, r, w commands. -u, --unbuffered: reduce buffering and flush output more often. -z, --null-data: use NUL ( \0) as line delimiter (useful for binary data).

Addressing (matching conditions)

/regexp/

: regular‑expression match. n (line number): match a specific line; ranges can be expressed as 1-$ (first to last line). addr1,addr2: match a range from addr1 to addr2 (both may be line numbers or regexes). addr1,+n: match addr1 and the following n lines. n~step: match every step ‑th line starting at line n (e.g., 1~2 for odd lines).

Delimiters

Any character can be used as a delimiter in a s command. When the chosen delimiter appears inside the pattern it must be escaped.

sed 's\|foo\|bar\g' file.txt
sed 's:foo:bar:g' file.txt
sed 's/\/bin/\/usr\/bin/g' file.txt

Variable references

Use double quotes when the sed script contains shell variables so that the variable is expanded before sed sees the script.

foo="world"
echo "hello $foo" | sed "s/$foo/librarookie/"
# Output: hello librarookie

Core editing commands

! addr!command: apply command to lines that do not match addr.

p addr p: print the current line (often used with -n).

l addr l: print the line showing non‑printing characters.

= addr =: print the current line number.

a\<text> addr a\text: append text after the matched line.

i\<text> addr i\text: insert text before the matched line.

c\<text> addr c\text: replace the entire matched line with text.

d addr d: delete the matched line.

r <filename> addr r file: read the contents of file and insert after the matched line.

w <filename> addr w file: write the matched line to file.

q [exit‑code] addr q: quit processing immediately (optional exit code).

s/regexp/replace/[flags] : substitute the first (or all with g) occurrence of regexp with replace. Flags: g: global replacement. n (numeric): replace only the n ‑th occurrence on the line. i: case‑insensitive match.

Advanced commands

h

: copy pattern space to hold space. H: append pattern space to hold space. g: replace pattern space with hold space. G: append hold space to pattern space. x: exchange pattern space and hold space. n: read the next line into pattern space (replaces current line). N: read the next line and append it to pattern space (separated by a newline). :label, t label, b label: define a label and branch to it conditionally ( t) or unconditionally ( b). y/old/new/: transliterate characters (like tr).

Typical use‑case examples

Range selection

# Print lines 3 to 5
sed -n '3,5p' file.txt
# Print from a start pattern to an end pattern
sed -n '/start/,/end/p' file.txt
# Print until the first blank line
sed -n '1,/^$/p' file.txt
# Print odd‑numbered lines
sed -n '1~2p' file.txt

Substitution (replace) examples

Replace the first occurrence on each line: sed 's/foo/bar/' file.txt Replace only the third match on each line: sed 's/foo/bar/3' file.txt Global replace (all occurrences): sed 's/foo/bar/g' file.txt Case‑insensitive replace: sed 's/foo/bar/i' file.txt Write replaced lines to a new file:

sed 's/foo/bar/w output.txt' file.txt

Insert, append, change

# Append after lines containing "hello"
sed -i '/hello/a\this is a test line' file.txt
# Insert before lines containing "librarookie"
sed -i '/librarookie/i\this is a test line' file.txt
# Change an entire line containing "librarookie"
sed -i '/librarookie/c\this is a test line' file.txt

Delete operations

# Delete the whole file content (dangerous)
sed -i 'd' file.txt
# Delete line 2
sed -i '2d' file.txt
# Delete blank lines
sed -i '/^$/d' file.txt
# Delete comment lines starting with #
sed -i '/^#/d' file.txt

Script files ( -f )

A sed script file contains one command per line; comments start with #. Example script.sed:

# Replace "old_string" on the first line
1s/old_string/new_string/
# Replace "foo" with "bar" from line 3 to 5
3,5s/foo/bar/g
# Delete lines containing "delete_this_line"
/delete_this_line/d
# Insert a line before lines matching "insert_before"
/insert_before/i\
This is the new line to insert
# Append a line at the end of the file
$ a\
This is the new line at the end of the file

Run with sed -f script.sed file.txt. Adding -n prevents automatic printing.

Hold space and pattern space

The hold space allows temporary storage between cycles. Example: duplicate lines containing test at the end of the file. sed -e '/test/h' -e '$G' file.txt Exchange hold and pattern spaces:

sed -e '/test/h' -e '/check/x' file.txt

Printing odd/even lines

# Odd lines using n
sed -n 'p;n' test.txt
# Even lines using n
sed -n 'n;p' test.txt
# Odd lines using step
sed -n '1~2p' test.txt
# Even lines using step
sed -n '2~2p' test.txt

Quit command

# Stop processing after line 10
sed '10q' file.txt