sed & awk Command Cheatsheet

Q: When should I use sed vs awk?

Simple rule: simple find-and-replace, line deletion/insertion → use sed. Field analysis, conditional filtering, calculations, formatted output → use awk. Specifically: sed is great for 'do something on matching lines' (substitute, delete, print). awk excels at 'compute on columns within rows' (sum, average, conditional filtering, reformatting output). They can be combined: `command | sed '...' | awk '...'`.

Q: What's the difference between `sed -i` on macOS and Linux?

macOS uses BSD sed which requires a backup extension argument after `-i` (even empty), while Linux's GNU sed allows `-i` standalone. Cross-platform approach: `sed -i '' 's/foo/bar/g' file` (macOS) and `sed -i 's/foo/bar/g' file` (Linux). Best practice: use `sed -i.bak` to create a backup file — compatible with both systems.

Q: How do I use quotes in awk and reference shell variables?

The awk script body must be wrapped in single quotes: `awk '{print $1}'`. To reference shell variables inside awk, use `-v` to pass them: `awk -v var="$shell_var" '{print var, $1}'`. Inside awk, double quotes are used for string concatenation: `awk '{print $1 ":" $2}'`. Note: string comparison uses `==`, regex matching uses `~`.

Q: What's the difference between basic regex (BRE) and extended regex (ERE) in sed?

Basic regex (BRE, default): meta characters like `+`, `?`, `{`, `}`, `(`, `)` need backslash escaping to activate their special meaning, e.g., `\+`, `$...$`. Extended regex (ERE, `sed -E`): these meta characters work without escaping. Always use `-E` for better readability. GNU sed also supports `-r`, but `-E` is more portable (also supported by BSD sed).

Q: What's the difference between NR and FNR in awk?

NR (Number of Records) is the global line counter, accumulating across all files. FNR (File-specific Number of Records) resets to 1 for each new file. Useful when processing multiple files: `awk 'FNR==NR {a[$1]; next} $1 in a' file1 file2` implements file matching/join by field.

Q: What special features does GNU awk (gawk) have?

GNU awk (gawk) adds: 1) Built-in functions like `asort()`, `asorti()` (array sorting), `gensub()` (global substitution with groups), `strftime()` (time formatting). 2) Network support `/inet/tcp/0/host/port`. 3) Co-process `|&` two-way pipe. 4) Profiler `--profile`. 5) Flexible `FPAT` for pattern-based field definition. Use `--posix` mode for strict compatibility with traditional awk.

sed (stream editor) and awk (text analysis tool) are the two powerhouses of Linux/Unix text processing. sed excels at pattern-based line operations and substitutions; awk shines at field analysis on structured text and report generation. This cheatsheet covers the most commonly used sed and awk commands — from basics to advanced techniques — each with practical examples.

Updated: 2026-07-16·38 commands

Quick Reference

Task

Command

Find and replace in a file	`sed -i 's/old/new/g' file`
Print lines 5-10	`sed -n '5,10p' file`
Remove blank and comment lines	`sed '/^#/d;/^$/d' file`
Print 1st and 3rd column of CSV	`awk -F',' '{print $1,$3}' file`
Sum a numeric column	`awk '{sum+=$1} END {print sum}' file`
Filter rows where column > 100	`awk '$NF > 100' file`
Formatted table output	`awk '{printf "%-20s %5d\\n", $1, $NF}'`

sed Basic Operations

``bash # Text substitution sed 's/pattern/replacement/' file # Replace first match per line sed 's/pattern/replacement/g' file # Global replace (all matches) sed 's/pattern/replacement/2' file # Replace only the 2nd match per line sed -i 's/pattern/replacement/g' file # In-place edit (GNU sed) sed -i '' 's/pattern/replacement/g' file # In-place edit (macOS BSD sed) sed -i.bak 's/pattern/replacement/g' file # Backup then replace (cross-platform)

# Line operations sed -n '5,10p' file # Print lines 5 through 10 sed -n '/pattern/p' file # Print lines matching pattern sed '5d' file # Delete line 5 sed '/pattern/d' file # Delete matching lines sed '$d' file # Delete the last line sed -n '1~2p' file # Print odd lines (every 2 lines starting at 1) sed '2~2d' file # Delete even lines

# Line editing sed '3a\\new line' file # Append a line after line 3 sed '2i\\new line' file # Insert a line before line 2 sed '5c\\replaced line' file # Replace line 5 with new text sed '/pattern/r insert.txt' file # Insert file content after matched line sed '/pattern/w output.txt' file # Write matching lines to a new file`

`sed Advanced Operations`

`bash # Range selection and compound operations sed -n '/start/,/end/p' file # Print lines between two patterns sed '/start/,/end/d' file # Delete lines within a range sed -n '5,/pattern/p' file # From line 5 to first pattern match sed '0,/pattern/{s/old/new/}' file # Replace only the first match sed '/pattern/!d' file # Delete non-matching lines (keep matches)

# Multiple commands sed '/pattern/{s/foo/bar/;s/baz/qux/}' file # Execute multiple subs on matched lines sed -e 's/foo/bar/' -e 's/baz/qux/' file # Multiple substitute commands sed 's/foo/bar/g; s/baz/qux/g' file # Semicolon-separated commands (GNU sed)

# Transformation and transliteration sed 's/.*/\\U&/' file # Uppercase entire line sed 's/.*/\\L&/' file # Lowercase entire line sed 's/\\b./\\U&/g' file # Capitalize first letter of each word sed 'y/abc/ABC/' file # Character mapping (a→A, b→B, c→C)

# Using the hold space sed -n 'h;n;G;p' file # Merge even lines into odd lines sed '1!G;h;$!d' file # Reverse file order (simulate tac)`

`sed Regex & Patterns`

`bash # Basic regex (BRE) sed '/^$/d' file # Delete empty lines (^ start, $ end) sed '/^#/d' file # Delete comment lines starting with # sed '/^[[:space:]]*$/d' file # Delete whitespace-only lines sed '/foo.*bar/d' file # Delete lines containing foo followed by bar sed 's/\\[.*\\]//g' file # Remove brackets and their contents sed 's/^[[:space:]]*//' file # Strip leading whitespace sed 's/[[:space:]]*$//' file # Strip trailing whitespace

# Extended regex (ERE) - use -E flag sed -E 's/([0-9]+)/[\\1]/g' file # Wrap numbers in brackets sed -E 's/^([^,]+),([^,]+)/\\2,\\1/' file # Swap first two fields (CSV) sed -E '/^(#|$)/d' file # Delete comments and empty lines sed -E 's/[[:space:]]+/ /g' file # Collapse multiple spaces into one sed -E 's|https?://([^/]+).*|URL: \\1|' file # Extract domain from URL

# Grouping and backreferences sed -E 's/([a-z]+)@([a-z]+)/\\1 AT \\2/g' file # Obfuscate email format sed -E 's/(.)(.)/\\2\\1/g' file # Swap adjacent characters sed -E 's/^(.{50}).*/\\1.../' file # Truncate long lines to 50 chars`

`awk Basic Operations`

`bash # Field printing awk '{print $1}' file # Print first field (default: space-delimited) awk '{print $0}' file # Print the entire line awk '{print $1, $3}' file # Print fields 1 and 3 awk '{print NR, $0}' file # Print with line numbers awk '{print NF, $0}' file # Show field count per line

# Custom delimiters awk -F',' '{print $1, $3}' file # Comma-separated (CSV processing) awk -F: '{print $1, $6}' /etc/passwd # Colon-separated awk -F'[ ,;]' '{print $1, $2}' file # Multiple delimiters awk 'BEGIN {FS=":"} {print $1}' file # Set delimiter in BEGIN block awk -F'\t' '{print $2}' file # Tab-separated

# Pattern matching awk '/pattern/ {print}' file # Print matching lines awk '/^error/ {print NR, $0}' file # Print lines starting with error + line numbers awk '!/pattern/ {print}' file # Print non-matching lines awk 'NR > 1 {print}' file # Skip header row awk 'NR % 2 == 0 {print}' file # Print even-numbered lines

# Conditional filtering awk '$3 > 100 {print}' file # Third column > 100 awk '$1 == "OK" {print}' file # First column equals OK awk 'length($0) > 80 {print}' file # Lines longer than 80 characters awk '$NF ~ /error/ {print}' file # Last column contains "error" awk '$1 >= 10 && $1 <= 100' file # Value between 10 and 100`

`awk Advanced Operations`

`bash # Statistics and aggregation awk '{sum+=$1} END {print "Sum:", sum}' file # Sum a column awk '{sum+=$1; count++} END {print "Avg:", sum/count}' file # Average awk '{if ($1 > max) max=$1} END {print "Max:", max}' file # Maximum awk 'NR==1{min=$1} {if ($1 < min) min=$1} END {print "Min:", min}' file # Minimum

# BEGIN / END blocks awk 'BEGIN {print "Report Start"} {print $0} END {print NR, "lines total"}' file awk 'BEGIN {FS=":"; OFS="--"} {print $1, $3}' /etc/passwd awk 'BEGIN {printf "%-20s %10s\\n", "Name", "Value"; print "------------------------"} {printf "%-20s %10d\\n", $1, $2}' file

# Arrays and counting awk '{a[$1]++} END {for (k in a) print k, a[k]}' file # Count value occurrences awk '{a[$1]+=$2} END {for (k in a) print k, a[k]}' file # Group by first column, sum second awk '!seen[$0]++' file # Remove duplicates (keep first occurrence)

# Multi-file processing awk 'FNR==NR {a[$1]; next} $1 in a' file1 file2 # File intersection (by field) awk 'FNR==NR {a[$1]=$2; next} {print $1, a[$1]}' map.txt data.txt # Key-based replacement awk 'NR==FNR {a[$1]++; next} !a[$1]' all.txt exclude.txt # Exclude items listed in another file

# Shell variables and external commands awk -v threshold=100 '$1 > threshold {print}' file # Pass shell variable awk -v name="$USER" '{print name, $0}' file # Pass environment variable awk '{"date" | getline now; print now, $0}' file # Execute external command awk '{system("ls -l " $1)}' file # Run system command per line

# String functions awk '{print toupper($0)}' file # Uppercase awk '{print tolower($0)}' file # Lowercase awk '{print substr($1, 1, 3)}' file # Substring (first 3 chars) awk '{print length($0), $0}' file # Show line length awk '{gsub(/old/, "new"); print}' file # Global in-line substitution`

`awk Field Processing`

`bash # Field selection and restructuring awk '{print $1, $(NF-1)}' file # Print first and second-to-last columns awk '{print $NF}' file # Print the last column awk '{$1=""; print $0}' file # Remove the first column awk '{$NF=""; $1=""; print $0}' file # Remove first and last columns awk -v n=3 '{print $n}' file # Dynamic column selection by variable

# Formatted output awk '{printf "%-20s %5d %8.2f\\n", $1, $2, $3}' file # Formatted table output awk 'BEGIN{OFS="|"} {print $1, $2, $3}' file # Set output delimiter to | awk '{print $1 "," $2 "," $3}' file # Manual CSV concatenation awk -v OFS='\t' '{print $1, $2}' file # Tab-separated output

# Field conditions awk '{if ($NF > 100) print}' file # Last column > 100 awk '$1 ~ /^[A-Z]/ {print}' file # First column starts with uppercase awk '{max=($1>$2)?$1:$2; print max}' file # Compare two columns, print larger awk 'NR==1{min=$1} {if ($1 < min) min=$1} END {print "Min:", min}' file # Track minimum value

# Advanced field parsing awk 'BEGIN {FPAT="[^,]*|\\"[^\\"]*\\""} {print $1}' file # Handle quoted CSV fields awk -F'[ ]+' '{print $1, $(NF-1)}' file # Variable-width space delimiter awk '{$1=$1; print}' file # Reformat line (clean up extra whitespace)``

sed Basics(7)

Command	Level	Example
`sed 's/pattern/replacement/' file` Replace the first occurrence of a pattern per line	Basic	`sed 's/foo/bar/' file.txt`
`sed 's/pattern/replacement/g' file` Replace all occurrences of a pattern globally	Basic	`sed 's/foo/bar/g' file.txt`
`sed -i 's/pattern/replacement/g' file` Edit file in-place with substitution	Basic	`sed -i.bak 's/oldtext/newtext/g' config.ini`
`sed -n '5,10p' file` Print lines from 5 to 10	Basic	`sed -n '10,20p' /var/log/syslog`
`sed '/pattern/d' file` Delete lines matching a pattern	Basic	`sed '/^$/d' file.txt`
`sed '5d' file` Delete a specific line by number	Basic	`sed '10d' file.txt`
`sed -n '/pattern/p' file` Print only lines matching a pattern	Basic	`sed -n '/ERROR/p' /var/log/app.log`

sed Advanced(6)

Command	Level	Example
`sed -n '/start/,/end/p' file` Print lines between two patterns (inclusive)	Intermediate	`sed -n '/BEGIN/,/END/p' document.txt`
`sed 's/.*/\U&/' file` Convert the entire line to uppercase	Intermediate	`sed 's/.*/\U&/' file.txt`
`sed -n '1~2p' file` Print odd-numbered lines (step pattern)	Intermediate	`sed -n '0~2p' file.txt`
`sed '/pattern/{s/foo/bar/;s/baz/qux/}' file` Execute multiple substitutions on matched lines	Intermediate	`sed '/user/{s/name/id/;s/email/mail/}' config.txt`
`sed '1!G;h;$!d' file` Reverse file order using the hold space	Expert	`sed '1!G;h;$!d' file.txt`
`sed 'y/abc/ABC/' file` Transliterate characters (character mapping)	Intermediate	`sed 'y/abcdefghijklmnopqrstuvwxyz/ABCDEFGHIJKLMNOPQRSTUVWXYZ/' file.txt`

sed Regex(5)

Command	Level	Example
`sed -E 's/([0-9]+)/[\1]/g' file` Extended regex with backreferences — wrap all numbers in brackets	Intermediate	`sed -E 's/([0-9]+\.[0-9]+)/[\1]/g' data.txt`
`sed -E '/^(#\|$)/d' file` Delete comment lines and empty lines	Intermediate	`sed -E '/^(#\|$)/d' config.conf`
`sed -E 's/[[:space:]]+/ /g' file` Collapse consecutive whitespace into a single space	Intermediate	`sed -E 's/[[:space:]]+/ /g' file.txt`
`sed -E 's/^([^,]+),([^,]+)/\2,\1/' file` Swap the first two fields in a CSV	Intermediate	`sed -E 's/^([^,]+),([^,]+)/\2,\1/' data.csv`
`sed 's/^[[:space:]]//; s/[[:space:]]$//' file` Strip both leading and trailing whitespace	Basic	`sed 's/^[[:space:]]//; s/[[:space:]]$//' file.txt`

awk Basics(7)

Command	Level	Example
`awk '{print $1}' file` Print the first field (default space-delimited)	Basic	`awk '{print $1}' access.log`
`awk '{print NR, $0}' file` Print each line prefixed with its line number	Basic	`awk '{print NR, $0}' file.txt`
`awk -F',' '{print $1, $3}' file` Use comma delimiter, print fields 1 and 3	Basic	`awk -F',' '{print $1, $3}' data.csv`
`awk '/pattern/ {print}' file` Print lines matching a pattern	Basic	`awk '/ERROR/ {print}' /var/log/app.log`
`awk 'NR > 1 {print}' file` Skip the first line (header row)	Basic	`awk 'NR > 1 {print}' data.csv`
`awk '$3 > 100 {print}' file` Print lines where the third column exceeds 100	Basic	`awk '$3 > 100 {print}' report.txt`
`awk 'length($0) > 80 {print}' file` Print lines longer than 80 characters	Basic	`awk 'length($0) > 80 {print}' source.py`

awk Advanced(6)

Command	Level	Example
`awk '{sum+=$1} END {print sum}' file` Sum the values in the first column	Intermediate	`awk '{sum+=$1} END {print "Total:", sum}' sales.txt`
`awk '{a[$1]++} END {for (k in a) print k, a[k]}' file` Count occurrences of each unique value in the first column	Intermediate	`awk '{a[$1]++} END {for (k in a) print k, a[k]}' ip.log`
`awk '!seen[$0]++' file` Remove duplicate lines (keep first occurrence)	Intermediate	`awk '!seen[$0]++' file.txt`
`awk 'FNR==NR {a[$1]; next} $1 in a' file1 file2` Find file intersection by key field between two files	Expert	`awk 'FNR==NR {a[$1]; next} $1 in a' keys.txt data.txt`
`awk -v threshold=100 '$1 > threshold {print}' file` Pass a shell variable into an awk script	Intermediate	`awk -v limit=500 '$1 > limit {print}' data.txt`
`awk '{gsub(/old/, "new"); print}' file` Perform in-line global substitution with gsub()	Intermediate	`awk '{gsub(/[0-9]+/, "NUM"); print}' data.txt`

awk Fields(7)

Command	Level	Example
`awk '{print $NF}' file` Print the last field of each line	Basic	`awk '{print $NF}' /etc/passwd`
`awk '{print $1, $(NF-1)}' file` Print the first and second-to-last columns	Intermediate	`awk '{print $1, $(NF-1)}' data.txt`
`awk '{$1=""; print $0}' file` Remove the first column (set it to empty string)	Intermediate	`awk '{$1=""; print $0}' file.txt`
`awk '{printf "%-20s %5d\\n", $1, $2}' file` Format output as a right/left-aligned table	Intermediate	`awk '{printf "%-20s %8.2f\\n", $1, $2}' report.txt`
`awk 'BEGIN{OFS="\|"} {print $1, $2, $3}' file` Set output field separator (OFS), commonly used for formatted table output	Intermediate	`awk 'BEGIN{OFS="\|"} {print $1, $2, $3}' data.txt`
`awk 'BEGIN {FPAT="[^,]\|\"[^\"]\""} {print $1}' file` Use FPAT to handle quoted CSV fields with embedded commas	Expert	`awk 'BEGIN {FPAT="[^,]\|\"[^\"]\""} {print $1}' complex.csv`
`awk '{$1=$1; print}' file` Reformat a line by cleaning up extra whitespace	Basic	`awk '{$1=$1; print}' messy.txt`

FAQ

When should I use sed vs awk?

What's the difference between `sed -i` on macOS and Linux?

How do I use quotes in awk and reference shell variables?

What's the difference between basic regex (BRE) and extended regex (ERE) in sed?

What's the difference between NR and FNR in awk?

What special features does GNU awk (gawk) have?

This cheatsheet is compiled from official tool documentation. Last updated: 2026-07-16.