awk.md

related
cheat-sheet sed

Matching

Match lines

awk '$1 ~ /pattern/ { ... }' infile

Matching for Conditions

awk '{if($1 ~ /pattern) { ... }}' infile

Negative Pattern

!/abc/ { print }

Blocks

BEGIN { print("only run at start") }

{ print("run for each line") }

END { print("only run at end") }

When using multiple files

BEGINFILE {print "--- " FILENAME " ---"}
ENDFILE { }

Control Flow

if (condition) { print("do something") } else { print("something else") }

# Ternary operator
print(a == b ? "equal":"more than equal")

# Loops
for(i=0; i<100; i+=1) { print(i) }

next          # skip to next line
nextfile      # skip to next file
exit          # terminate 
exit(1)       # with exit code

Variables

Name	Meaning	Description
NF	number of fields	always givens the number of fields in current line based on IFS
ARGV	like in C	all extra CLI params not consumed by awk
ARGC	like in C	nr of extra CLI params not consumed by awk

Input/Output Control Variables

Name	Meaning	Description
ORS	Output Record Separator	String added after each print output (can be a regex).
RS	(Input) Record Separator	Decides how input lines will be split (can be a regex).
IFS	Input Field Separator	Decides how input fields will be split.
FS	Field Separator	Decides how output fieds will be joined.

You usually use those variables by either setting them in the BEGIN block or using -v.

Examples

awk -v ORS='\r\n' '{ print }'                # Use \r\n line endings instead of \n
find -print0 | awk -v RS='\0' '{ print }'    # Safe filename processing by using \0 as input separator
awk -v IFS=, '{ print $3 }' my.csv           # Process CSV by splitting fields on comma
awk 'BEGIN { FS="," } '{ print $1, $3 }'     # Produce CSV from whitespace separated input

Passing shell variables

Use -v <key>=<value> or ENVIRON["<variable>"]

awk -v HOME="$HOME" 'BEGIN { print(HOME) }'

awk 'BEGIN { print ENVIRON["SHELL"] }'

Advanced Use of awk

Negative Indizes

awk '{print $(NF - 1)}'