awk '{print $2}': print a specific column

awk '{print $2}' file prints the second column of every line. By default awk splits each line on runs of spaces and tabs, ignoring leading blanks, and numbers the fields $1, $2, ...; $0 is the whole line and $NF the last field. For other separators set -F, for example awk -F, '{print $2}' file.csv or awk -F: '{print $1}' /etc/passwd.

Examples

printf 'alice  admin   42\nbob    dev     37\ncarol  ops     29\n' > users.txt
echo '--- second column'
awk '{print $2}' users.txt
echo '--- first and third, separated by a comma'
awk '{print $1 "," $3}' users.txt
echo '--- last column'
awk '{print $NF}' users.txt
echo '--- columns with a header and a condition'
awk 'BEGIN {print "name age"} $3 > 30 {print $1, $3}' users.txt

--- second column
admin
dev
ops
--- first and third, separated by a comma
alice,42
bob,37
carol,29
--- last column
42
37
29
--- columns with a header and a condition
name age
alice 42
bob 37

print $1, $3 (with a comma) separates fields with the output field separator OFS, a single space by default; print $1 $3 (no comma) joins them with nothing in between.

Other separators

printf 'id,name,city\n1,Ada,London\n2,Linus,Helsinki\n' > people.csv
awk -F, '{print $2}' people.csv
echo '---'
awk -F, -v OFS='\t' 'NR > 1 {print $3, $2}' people.csv
echo '---'
echo 'root:x:0:0:root:/root:/bin/bash' | awk -F: '{print $1, $7}'

name
Ada
Linus
---
London	Ada
Helsinki	Linus
---
root /bin/bash

Common tasks

A range of columns, or a column by name

printf 'id name city country\n1 Ada London UK\n2 Linus Helsinki FI\n' > t.txt
echo '--- columns 2 to the end'
awk '{ for (i = 2; i <= NF; i++) printf "%s%s", $i, (i < NF ? OFS : ORS) }' t.txt
echo '--- the column named city'
awk -v col=city 'NR == 1 { for (i = 1; i <= NF; i++) if ($i == col) c = i; next } { print $c }' t.txt

--- columns 2 to the end
name city country
Ada London UK
Linus Helsinki FI
--- the column named city
London
Helsinki

Pitfalls

• Double quotes around the program. The shell expands $2 before awk runs; the program becomes {print } and every line is printed in full. Use single quotes.
• Windows line endings. A file saved with CRLF keeps a \r at the end of the last field, so $NF looks right on screen but comparisons such as $NF == "UK" fail. Strip it with tr -d '\r' or sub(/\r$/, "").
• Empty fields. With the default separator, runs of blanks count as one separator, so an empty column in an aligned report disappears and later columns shift left. Fixed-width data needs substr() or GNU awk's FIELDWIDTHS.
• Unquoted file names. A space in a path splits it into two arguments and awk stops with cannot open file for reading. For quoting mistakes inside the program, see awk syntax error.

To add up a column instead of printing it, see awk sum column; for separators in depth, awk -F.

cut -d, -f2 file is an alternative for a single-character delimiter, but cut treats each space as a separator, so it does not handle the aligned columns in users.txt the way awk does.

FAQ