Question

Given a string such as abc-def-ghi-jkl.
Can the sed command extract the substring to the left of the nth separator character such as abc for 1, abd-def for 2 etc?

I need it for making creating files based on a subset of the computers hostname, normally abc-def-xyz.domain.com.

Having looked around it appears that awk, cut and sed are combined in this kind of extractions but I prefer something that can be used as part of a command line.

Answer 1

cutis your friend:

$ echo "abc-def-ghi-jkl" | cut -d- -f1
abc
$ echo "abc-def-ghi-jkl" | cut -d- -f-2
abc-def
$ echo "abc-def-ghi-jkl" | cut -d- -f-3
abc-def-ghi
$ echo "abc-def-ghi-jkl" | cut -d- -f-4
abc-def-ghi-jkl

other possibilities:

$ echo "abc-def-ghi-jkl" | cut -d- -f2
def
$ echo "abc-def-ghi-jkl" | cut -d- -f3
ghi
$ echo "abc-def-ghi-jkl" | cut -d- -f4
jkl