sed

Use

printf 'hello makers\n' | sed 's/makers/world/'

What It Does

sed transforms text as it streams through editing rules. Most beginner use is substitution.

Simple Substitutions

Replace the first match on each line:

printf 'hello makers\n' | sed 's/makers/world/'

Replace every match on each line:

printf 'ha ha ha\n' | sed 's/ha/HA/g'

Use a different delimiter when the pattern contains /:

printf 'https://example.org\n' | sed 's#https://#http://#'

Delete blank lines:

sed '/^$/d' notes.txt

Print only lines 1 through 5:

sed -n '1,5p' notes.txt

Capture Example

printf '# heading\n' | sed 's/^# \(.*\)$/<h1>\1<\/h1>/'

• s/old/new/ means substitute.
• ^ anchors the start of the line.
• $ anchors the end.
• \(.*\) captures text in basic sed regex.
• \1 reuses the captured text.
• <\/h1> escapes / because / is the delimiter.

In-Place Editing

Use in-place editing only after testing without it:

cp page.html page.html.backup
sed 's/old/new/' page.html
sed -i 's/old/new/' page.html

sed -i changes the file. Test on stdout first.

Watch Out

• Keep sed scripts in single quotes so Bash does not alter them.
• If output is unchanged, the pattern did not match.
• If \1 fails, you did not create a capture group in sed’s regex dialect.

Docs Pointers

• Run man sed.
• Read GNU sed manual.
• Read regular expressions, text transforms, quoting, and awk.