bash Cookbook

Isolating Specific Fields in Data

Problem

You need to extract one or more fields from each line of output.

Solution

Use cut if there are delimiters you can easily pick out, even if they are different for the beginning and end of the field you need:

# Here's an easy one, what users, home directories and shells do
# we have on this NetBSD system
$ cut -d':' -f1,6,7 /etc/passwd
root:/root:/bin/csh
toor:/root:/bin/sh
daemon:/:/sbin/nologin
operator:/usr/guest/operator:/sbin/nologin
bin:/:/sbin/nologin
games:/usr/games:/sbin/nologin
postfix:/var/spool/postfix:/sbin/nologin
named:/var/chroot/named:/sbin/nologin
ntpd:/var/chroot/ntpd:/sbin/nologin
sshd:/var/chroot/sshd:/sbin/nologin
smmsp:/nonexistent:/sbin/nologin
uucp:/var/spool/uucppublic:/usr/libexec/uucp/uucico
nobody:/nonexistent:/sbin/nologin
jp:/home/jp:/usr/pkg/bin/bash

# What is the most popular shell on the system?
$ cut -d':' -f7 /etc/passwd | sort | uniq -c | sort -rn
10 /sbin/nologin
2 /usr/pkg/bin/bash
1 /bin/csh
1 /bin/sh
1 /usr/libexec/uucp/uucico

# Now let's see the first two directory levels
$ cut -d':' -f6 /etc/passwd | cut -d'/' -f1-3 | sort -u
/
/home/jp
/nonexistent
/root
/usr/games
/usr/guest
/var/chroot
/var/spool

Use awk to split on multiples of whitespace, or if you need to rearrange the order of the output fields. Note the → denotes a tab character in the output. The default is space but you can change that using $OFS:

# Users, home directories and shells, but swap the last two # and use a tab delimiter $ awk 'BEGIN ...