CODEWITHSUNDEEP
parsingawkgrep
mcv110
mcv110

Reputation: 13

how to parse multiple strings from a linux file

I have a script that that I want to parse the username and IP address from.

Here is a sample of the script. this entry exists for all users, about 2000 lines in the script.

 if [ "$common_name" = "NUMERICUSERNAME" ]; then
  if [ $(/usr/sbin/iptables -t nat -L -v | grep to: | cut -d : -f 2) = "IP.AD.DR.ESS" ] ; then
   /usr/sbin/iptables -t nat -D POSTROUTING -s $(/usr/sbin/iptables -t nat -L -v | grep IP.AD.DR.ESS | gawk '{ print $(NF-2) }') -j SNAT --to-source IP.AD.DR.ESS ;
  fi
  /usr/sbin/iptables -t nat -A POSTROUTING -s $ifconfig_pool_remote_ip -j SNAT --to-source IP.AD.DR.ESS
  fi

The username is a 15 digit numeric string.

I have tried parsing with awk...something like awk '{print $5,$25}' filename but cant find the correct fields to capture.

I have also tried different forms of grep but do not know how to capture BOTH fields without causing the data to misalign. for example grep -oP '[0-9-]{10} [0-9:]{8}' filename gives me all the IPs or all the usernames...but I cant seem to figure out how to get both.

What is the most efficient way to get the output of NUMERICUSERNAME and is matching IP.AD.DR.ESS. something like this

NUMERICUSERNAME1 IP.AD.DR.ESS1
NUMERICUSERNAME2 IP.AD.DR.ESS2
....
NUMERICUSERNAME2000 IP.AD.DR.ESS2000

Also note that the entry has the IP.AD.DR.ESS in more than one location and that caused me to capture duplicates. can that be avoided as well?

EDIT: output of

$ awk '/common_name/ {u = substr($5,2,15)}
       /--to-source/ {print u, $(NF-1)}' < script

output:

 IP.AD.DR.ESS1
 --to-source
 IP.AD.DR.ESS10
 --to-source
 IP.AD.DR.ESS100
 ...
 IP.AD.DR.ESS75
 --to-source
NUMERICUSERNAME IP.AD.DR.ESS76
SAMENUMERICUSERNAME --to-source
SAMENUMERICUSERNAME IP.AD.DR.ESS77

notepad++ screen capture

EDIT

sample output from solution

NUMBERICUSERNAME1 IP.AD.DR.ESS1
NUMBERICUSERNAME1 --to-source
NUMBERICUSERNAME2 IP.AD.DR.ESS2
NUMBERICUSERNAME2 --to-source

Upvotes: 0

Views: 151

Answers (1)

Renaud Pacalet
Renaud Pacalet

Reputation: 29220

Something like this, maybe:

$ awk '/common_name/ {u = substr($5,2,15)}
       /iptables -t nat -A POSTROUTING/ {print u, $NF}' < script
NUMERICUSERNAME IP.AD.DR.ESS

If the /common_name/ and/or /iptables -t nat -A POSTROUTING/ regular expressions catch unwanted lines you can improve them a bit to make them more selective. Example:

/"\$common_name"\s*=\s*"[[:digit:]]{15}"/

Upvotes: 1

Related Questions