CIS241

System-Level Programming and Utilities

Pattern scanning and processing language (awk)

Erik Fredericks, frederer@gvsu.edu
Fall 2025

Based on material provided by Erin Carrier, Austin Ferguson, and Katherine Bowers

awk

Pattern scanning and processing language

• Data driven, not procedural
  • Describe data to work with
  • Tell it what to do when pattern found!

Good for data in columns

• gawk is the GNU implementation - what we'll use

gawk program

BEGIN { ... initialization gawk statements ... }
gawk comamnds to run on each line of the file
END { ... finalization gawk statements ...}

Each command has the form: pattern { action }

• action = one or more statements enclosed in braces
• Pattern can be regex
• No pattern - action performed on all lines

Fields

Each line is made up of fields

• Field separator distinguished fields

  • Default - space
  • Change value of FS to use another
  • or use -F to change

• Reference field by $# ($0 is entire line)

• NF = number of fields on current line

• NR = record number of current line

Sample data

• Record is a row, field is a cell!

Running gawk

gawk [options] program [input_files]

gawk -f program_file [input_files]

Can also create a script telling it to run with gawk instead!

• #!/usr/bin/env gawk -f
  • Also can use path directly to gawk
  • Script needs to be executable!

Patterns

Pattern can be regex (/regex/)

• ~ used for matching regex
• !~ tests for not matching regex

Pattern can also compare field or variable to value:

• ==, !=, <, <=, >, >=

BEGIN and END are special patterns

Nothing for pattern -> applies to all records

Can combine patterns with && (and), || (or)

Variables

Can hold strings/numeric values

Typically initialized in BEGIN

Default -> initialized to empty string or 0

Standard arithmetic operators available

https://medium.com/@redswitches/the-gawk-command-in-linux-with-10-examples-092900b06ca5

gawk '/[e,E]$/ {print}' words.txt