Shell Programming and Scripting

You are on your way. But what is inside \1 ? How did I collect the pattern inside \1 ?

I thought that is what's inside the (, ) but that does not encompasses the whole pattern.

Ok, I just moved the first \( and now \1 is the whole pattern, I can replace it and it works Thanks for your help so far.

Now I need to do something else. I actually need to replace the IP with a hostname. I have a small script that does this but I am not sure how to pass the \1 to my script, inside the sed statement.

Normally the script does this:

Code:

server:~$ res 10.1.1.3
host1.domain.com

The res script is just a simple grep script:

Code:

server:~$ cat bin/res
#!/usr/bin/bash
grep -w $1 /etc/hosts | cut -f2

But when I try to integrate it into the sed line, I get this:

Code:

server:~$ echo "This is an IP: 10.10.2.10" | sed -e "s/\([0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\)/`res "\1" `/g"
grep: RE error in \<\1\>: number in \[0-9] invalid
This is an IP:

The grep inside "res" does not like the \1, which I kinda understand. Thing is, if I do a simple echo "\1" instead of the res "\1" then it works, the \1 is replaced with the actual IP from the sed statement.

Any ideas ?

Thanks,
Sylaan

What about

Code:

NAME=$(res 10.1.1.3)
echo "This is an IP: 10.10.2.10" | sed -e "s/[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}\.[0-9]\{1,3\}/$NAME/g"

That works but I kinda need to use the \1, the IP address that I matched. The lines I am going to parse will be syslog entries, which contain IP addresses. I just need to tail the syslog, pipe it through the sed statement and replace the IP address with its equivalent hostname (which I get via the small res script).