Extract IP addresses in each line of file in shell

Question

I have this input file in ubuntu:

146.14.142.96.17747 197.102.40.184.13748:
146.14.142.96.17747 197.102.40.184.13749: 
146.14.142.96.17747 197.102.40.184.13750:
146.114.142.96.17747 197.102.40.184.13751:
46.14.142.96.17747 197.102.40.184.13752:

and I'd like to have the output like this using shell scripting:

separate two IPs without ports number I mean delete ports

146.14.142.96 197.102.40.184
146.14.142.96 197.102.40.184 
146.14.142.96 197.102.40.184
146.114.142.96 197.102.40.184
46.14.142.96 197.102.40.184

Answer 1

For lines formatted exactly as shown in the question, this will do:

sed -E 's/\.[0-9]+[ :]/ /g' input-file

How it works:

• The -E switch enables Extended Regular Expressions.

• s/SEARCH/REPLACE/g globally (/g) replaces (s/) SEARCH with REPLACE.

• \.[0-9]+[ :] matches a dot following any positive number of digits following a space or a colon.

However, this will break if the formatting varies even slightly. This approach may result robuster:

sed -E 's/(([0-9]+\.){3}[0-9]+)[^ ]+/\1/g' input-file

How it works:

• ([0-9]+\.){3}[0-9]+ matches an IP (three digit groups followed by dots plus an additional digit group).

• The surrounding parentheses declare the previous match as the first submatch (\1).

• [^ ]+ matches any non-space character that follows the IP.

Answer 2

Do a search and replace using regex

(\d+\.\d+\.\d+\.\d+)\.\d+(:?)

and replace text

\1

Various tools support regex search and replace though the dialect can be slightly different. The above works with Notepad++.

Or in vim you can do

:s/\(\d\+\.\d\+\.\d\+\.\d\+\)\.\d\+\(:\?\)/\1/g