Getting delimited substrings using sed+regexp

Question

I'm trying to get substrings using sed with regexp. I want to get the first and second "fields" delimited by ":".

To get the first field I used the following command, but don't know how to get the second field.

Command used to get the first field:

sed -r -n '1,2 s/([^:]+).*/\1/p' /etc/passwd

Input file (example):

root:x:0:0:root:/root:/bin/bash   
daemon:x:1:1:daemon:/usr/sbin:/usr/sbin/nologin

Command's result:

root   
daemon

But I tried do get the first ("root") and second ("x") fields (examples based on the file's first line only), but I did't succedded.

I tried:

sed -r -n '1,2 s/([^:]+).*([^:]+).*/\1 \2/p' /etc/passwd

Command's result:

root h   
daemon n

Desired result:

root x   
daemon x

With awk: `awk -F ':' '{print $1,$2}' file` – Cyrus Jul 01 '19 at 03:22 — Cyrus, Jul 01 '19 at 03:22

Yang · Accepted Answer · 2019-07-01T04:47:59.040

0

sed uses greedy match. In

sed -r -n '1,2 s/([^:]+).*([^:]+).*/\1 \2/p' /etc/passwd
                        ^^

.* matches as many characters as possible. You need

sed -r -n '1,2 s/([^:]+):([^:]+).*/\1 \2/p' /etc/passwd
                        ^

Demo: http://ideone.com/wjL7Za.

By the way, a simpler way to do this is using cut:

cut -d ":" -f 1,2 --output-delimiter=' ' /etc/passwd

Demo: http://ideone.com/stJdSy.

edited Jul 01 '19 at 04:47

answered Jun 30 '19 at 23:59

Yang

Fantastic! Great explanation. And the websites' links are amazing. Thanks – Edilson Marcussi Jul 03 '19 at 00:05

score 0 · Answer 2 · answered Jul 01 '19 at 02:50

0

Another expression that would return the desire result would be:

([a-z]+):([a-z]+).*

sed -r -n '1,2 s/([^:]+):([^:]+).*/\1 \2/p'

answered Jul 01 '19 at 02:50

Emma

2 Answers2