cut-vs-awk.md

cut vs awk behavior

Following example compares cut strict delimiting behavior to more flexible awk one.

Is awk really needed when cut is already out there?

cut utility is good enough to be used instead of awk when we work with fixed text format like csv or similar...

$ head -5 /etc/group
root:x:0:
bin:x:1:
daemon:x:2:
sys:x:3:
adm:x:4:

$ head -5 /etc/group | cut -d : -f 3
0
1
2
3
4

$ head -5 /etc/group | awk -F: '{print $3}'
0
1
2
3
4

cut limitations

When looking more deep we found that cut defines delimiter as single character and thus treat sequence of delimiter characters as multiple delimiters (unlike awk) which makes troubles when parsing text formated tables ...

$ ps auxwww | head -5
USER       PID %CPU %MEM    VSZ   RSS TTY      STAT START   TIME COMMAND
root         1  0.2  0.1 194336  8816 ?        Ss   17:09   0:01 /usr/lib/systemd/systemd --switched-root --system --deserialize 21
root         2  0.0  0.0      0     0 ?        S    17:09   0:00 [kthreadd]
root         3  0.0  0.0      0     0 ?        S    17:09   0:00 [ksoftirqd/0]
root         5  0.0  0.0      0     0 ?        S<   17:09   0:00 [kworker/0:0H]
$ ps auxwww | head -5 | cut -d " " -f 2





$ ps auxwww | head -5 | awk '{print $2}'
PID
1
2
3
5

Summary

• cut can replace awk in narrow set of use-cases only