Ubuntu – Capturing only a section of a command output

bashcommand linescripts

If I were to type a certain command, it would give me a long output in the terminal. I only need a few characters of this output. How do I, in a bash script, save only certain characters from a terminal output as a variable?

For example, if my script is:

#!/bin/bash

ifconfig

I could save the whole output as a variable using

OUTPUT="$(ifconfig)"

But I only want my IP address copied, which is after inet [IP ADDRESS]

How would I go about doing this?

Edit: IP Address is just being used here as an example as I cannnot mention the command on here for security reasons, it's actually 'phrase: [five digit number]' in amongst other text and I just need that five digit number. Sorry for any confusion

Best Answer

Using `grep`

ifconfig | grep -oP '(?<=inet addr:)[\d.]+'

This uses grep's Perl-style regular expressions to select the IP address that follows the string inet.

So, to save that in a variable, just put both commands inside the $():

output=$(ifconfig | grep -oP '(?<=inet addr:)[\d.]+')

The above will save the IP addresses for all active interfaces on your system. If you just want to save the output for one interface, say eth0, then use:

output=$(ifconfig eth0 | grep -oP '(?<=inet addr:)[\d.]+')

Using `awk`

ifconfig eth0 | awk -F'[ :]+' '/inet /{print $3}'

/inet / selects lines that contain inet. On those lines, the third field is printed where a field separator is an y combination of spaces or colons.

Using `sed`

ifconfig eth0 | sed -En 's/.*inet addr:([[:digit:].]+).*/\1/p'

Alternate `ifconfig`

There is another version of ifconfig which produces output like inet 1.2.3.4 instead of inet addr:1.2.3.4. For that version, try:

ifconfig | grep -oP '(?<=inet )[\d.]+'

Or:

ifconfig eth0 | awk '/inet /{print $2}'

Or, use this sed command which works with either version of ifconfig:

ifconfig eth0 | sed -En 's/.*inet (addr:)?([[:digit:].]+).*/\2/p'

Related Solutions

Bash Scripts – Differences Between Terminal Command-Line and Bash Script Environments

Your question cannot be answered quite generally. Some comments though. On my version of bash (3.1.17(1)-release), your command doesn't have the desired output even when run from the command line; the same with zsh. So presumably something is fishy about your command. I don't know what "##+(0)" is suppoed to accomplish, but "#0" succeeds in removing one leading zero. This show a way to remove arbitrary numbers of zeros.

If there really is a difference between the behavior on the command line and from a script, then most likely the script uses a different interpreter (different bash version, bash instead of zsh) or different shell options (try running shopt). The latter difference might be the result of your interactive shell sourcing $HOME/.bashrc and $HOME/.profile whereas scripts generally don't. That shouldn't affect environment variables, as they're inherited if exported, but it should affect shell options, which need to be set in every shell.

Syslog – How to Use Syslog for Else Output with Whois Query

Yet another solution:

awk '{ 
      for (i = 1; i <= NF; i++) 
        if ($i ~ /^SRC=/) 
          print substr($i, 5) 
    }' /var/log/syslog |
  sort -u |
  while read ip; do
    printf ' === %s ===\n' "$ip"
    whois "$ip"
  done

If you only want to select lines of syslog containing the string INVALID STATE, then the above code can be modified as follows

awk '/INVALID STATE/ { 
      for (i = 1; i <= NF; i++) 
        if ($i ~ /^SRC=/) 
          print substr($i, 5) 
    }' /var/log/syslog |
  sort -u |
  while read ip; do
    printf ' === %s ===\n' "$ip"
    whois "$ip"
  done

Best Answer

Using grep

Using awk

Using sed

Alternate ifconfig