CODEWITHSUNDEEP
regexbashsed
Christian M
Christian M

Reputation: 70

Isolate part of string with sed

I'm currently making a script in Bash to make my life slightly easier at work. I'm trying to isolate the host name from my nslookup call, but I'm having some issues.

The output from nslookup is 1.0.0.10.in-addr.arpa name = nuc3.homenet.no and I'm trying to match and save only the nuc3 part of the string and save it in a variable, by capturing nuc3 and replacing the whole string with the captured group.

NSLOOKUP=$(nslookup 10.0.0.1) 
NUC=$(echo ${NSLOOKUP} | sed -E 's/.*?(nuc\d+).*/\1/'

\d+ is predominantly because there are several machines named the same way with numbers ranging from 1-10.

The output from sed would just default to printing the original string, so I'd something sketchy is going on. I've successfully used the same method to isolate strings earlier with no problems.

Upvotes: 1

Views: 130

Answers (2)

anubhava
anubhava

Reputation: 786091

sed doesn't support \d. Just use [0-9] and also match a space before nuc to avoid false matches:

echo '1.0.0.10.in-addr.arpa name = nuc3.homenet.no' | sed -E 's/.* (nuc[0-9]+).*/\1/'


nuc3

Upvotes: 1

Matias Barrios
Matias Barrios

Reputation: 5054

Why not this ?

echo '1.0.0.10.in-addr.arpa name = nuc3.homenet.no' | awk -F'=' '{print $2}' | tr -d -c '[a-zA-Z0-9.\n]'

Or also this :

echo '1.0.0.10.in-addr.arpa name = nuc3.homenet.no' | grep -o -P  '[a-zA-Z]+[0-9]+[.][a-zA-Z]+[.][a-zA-Z]+'

Upvotes: 1

Related Questions