I got this script from squid-cache.org. This onverts a fieldin access log to a readable format.

cat access.log | perl -nwe 's/^(\d+)/localtime($1)/e; print' > test

i want to modify this script so that it also converts the second field which is the time elapsed in millyseconds to minutes or seconds. how can i do that? I don't know perl.

Anyway, this snippet of perl takes the first series of digits in a line and converts it to the ctime value. You didn't post the format of the logfile, however assuming after the first field there are spaces before the second field, subsitituting the regex for something like this might do the trick:

s/^(\d+)\s+(\d+)/localtime($1) . " " . $2\/1000/e

What do you mean by seconds and miliseconds by the way? Do you want it formatted like XX mins xx s?

i tried with a command line as below

cat access.log | perl -nwe 's/^(\d+)\s+(\d+)/localtime($1) ." ". $2\1000/e; print';.


I got the following error

Can't use \1 to mean $1 in expression at -e line 1.
Backslash found where operator expected at -e line 1, near "$2\"
(Missing operator before \?)
syntax error at -e line 1, near "$2\"
Execution of -e aborted due to compilation errors.
./view_log2: line 1: .: filename argument required
.: usage: . filename

You may want to see the format of access.log
a single line is like this.

1090827471.120 2205 192.168.0.32 TCP_MISS/200 24692 GET http://mail.yahoo.com/? - DIRECT/216.109.127.60 text/html


The first field is the time of request in Unix format second field is the time elapsed. the two fields are to be converted.

I have tried it out, The first one gives the below o/p
Mon Aug 2 09:06:00 2004 0.705
The second one
Sun Aug 1 11:28:23 2004 1.038 192.168.0.32 TCP_MISS/200 197 GET http://toolbarqueries.google.com/search? - DIRECT/64.233.161.99 text/html

Thanks a lot.
Will u please help me to modify the Ist commandline sothat it also includes the third field too?

Another question,Can u suggest a good tutorial for perl?

For me the third field is the IP, what do you call third field ?

For my part, I learnt Perl on the net and with practice on the command line and scripts, so I suggest you to practice

[edit]
I think I understand what you want :

u r correct. the third field is ip. I want to trace the requests to a particular site
Your script will do that.
Thanks alot!

I have another question to ask.

I have to keep seperate log files for each PC. I put the script u have given in a file named view_log_minimal


#file view_log_minimal
grep "websitename" /var/log/squid/access.log | perl -p -nwe 's/^(\d+\.\d+)\s+(\d+)\s(\d+\.\d+\.\d+\.\d+)\s(.*$)/localtime($1)." ".($2\/1000)." ".($3)/e'

i wrote another shell script to get ip address wise output

#file ip_wise_list

#!/bin/bash

X=0
while [ $X -le 60 ]
do
X=$((X+1))
IPADR=$(printf "192.168.0.%d" $X)
./view_log_minimal | grep $IPADR
done


There is some problem. It takes more than one second to process each ip. It shows the result .

Can u tell me what is the problem with my script?

If u have an alternative with perl, i will be happy to implement that