topic parse help needed in Operating System - Linux

parse help needed

Chris H_3 — Mon, 03 May 2004 15:17:15 GMT

Hi all,
I have a router log file that I would like to parse. It contains 4 columns of IP's. I would like to sort them based on column 4, and also add a number to each row of how many occurances of that particular connection.
i.e
10.12.5.24 172.10.5.10 172.10.5.12 172.50.101.5
10.12.5.20 172.10.5.10 172.10.5.12 172.50.101.5
10.12.5.24 172.10.5.10 172.10.5.12 172.50.101.5
10.12.5.24 172.10.5.10 172.10.5.12 172.50.101.5
...
so there is one connection from 10.12.5.20 and 3 from 10.12.5.24, so the output would be

10.12.5.24 172.10.5.10 172.10.5.12 172.50.101.5 3
10.12.5.20 172.10.5.10 172.10.5.12 172.50.101.5 1
...
...

I started with an awk script, something like:
cat filename.txt | awk '{print $4" "$3" "$2" "$1} | sort | uniq

but it didn't look quite right, and I couldn't figure out how to generate the number of occurances, and my perl knowledge is about useless :(

Any help would be appreciated, and if more info is needed, let me know

Thanks!

CH

Re: parse help needed

Martin P.J. Zinser — Mon, 03 May 2004 16:29:58 GMT

Hi,

maybe not the greatest, but should do the trick :-)

$file = "temp.txt";

open(IN, $file);
while(){
chomp();
@words = split();
$srec = join(" ", $words[3], $words[0], $words[1], $words[2]);
$sarr{$srec} = $_;
$count{$_}++;
}
close(IN);

foreach $key (sort keys %sarr){
printf "%s %s\n", $sarr{$key}, $count{$sarr{$key}};
}

Hopefully ITRC does not mess this up completely.

Greetings, Martin

Re: parse help needed

Francisco J. Soler — Mon, 03 May 2004 16:37:32 GMT

Hi CH,

try this:

sort +4 -5 filename.txt | uniq -c | sort -r

and if you need the counter first:

sort +4 -5 filename.txt | uniq -c | sort -r | awk '{printf $2" "$3" "$4" "$5" "$1"\n"}'

Frank.

Re: parse help needed

Chris H_3 — Mon, 03 May 2004 17:47:29 GMT

Oooo, that sort is slick, thanks Frank.

Actually, they both did the trick, so thank you both. Cool to see it from two ways.

Thanks again!

CH