hisnewbness has asked for the wisdom of the Perl Monks concerning the following question:

Know you guys are busy, but if I may have a moment of your time perhaps someone can help out a new to perl guy.

Situation: I'm writing a simple script that parses a syslog file full of data that looks like this:

# Aug  5 11:50:50 ev-cis-cons2 142682254: 2004 Aug 05 11:50:50.040 EDT
+ -4:00 %AUTH-5-28: RPT=88653: 10.1.1.1: User [user0] Group [group0] d
+isconnected:  Duration: 1:16:40  Bytes xmt: 1441872  Bytes rcv: 56377
+6  Reason: User Requested
# Aug  5 11:51:52 ev-cis-cons2 142683959: 2004 Aug 05 11:51:52.760 EDT
+ -4:00 %AUTH-5-28: RPT=88654: 10.1.1.2: User [user1] Group [group1] d
+isconnected:  Duration: 3:55:38  Bytes xmt: 21588288  Bytes rcv: 2879
+544  Reason: Lost Service
# Aug  5 11:52:22 ev-cis-cons2 142685133: 2004 Aug 05 11:52:22.650 EDT
+ -4:00 %AUTH-5-28: RPT=88655: 10.1.1.3: User [user2] Group [group2] d
+isconnected:  Duration: 1:35:37  Bytes xmt: 1424368  Bytes rcv: 31430
+4  Reason: User Requested

[download]

(Reluctant editorial note from Ovid: Fixed this again because we were still getting complaints. The above log info is now in <code> tags because it was still not wrapping well for people. Just note that the '# ' at the beginning of each line was not in the original.)

My script:

#Look in daily syslog file and grab the
#user, group, connection duration, bytes xmt and bytes rcv
#Stick this data into a daily report in .csv format
#


#TIME & DATE STUFF
$stime = localtime;
@time1 = split(' ', $stime);
$mon = $time1[1];
$day = $time1[2];
$year = $time1[4];

$filename = "dailyreport-$mon-$day-$year.csv";

#Open the log file for reading

open (LOG, "</var/log/ciscon.log") || die "Couldn't open log file $!";
open (TMP, ">>tmpfile.txt") || die "Couldn't open tmp file $!";
open (DREPORT, ">>/root/reports/$filename") || die "Couldn't open csv 
+$!";
print DREPORT "DAILY LOG FOR $mon $day $year\n";
print DREPORT "\n";
print DREPORT "USERNAME,GROUP,DURATION,BYTES XMT,BYTES RCV,\n";
print DREPORT "\n";

#Pump data into array and parse for usefull data
#Usefull data goes into tmpfile

while( <LOG>)
        {
        @tmp = grep {/disconnected/} $_;
        print TMP "@tmp";
        }

#Close stuff we dont need

close TMP;
close LOG;

#Open temp file and stuff into array by spaces print data we
#want into the daily csv file

open (TMP1, "<tmpfile.txt")  || die "Can't read from tmpfile $!";
while( <TMP1> )
        {
        chomp $_;
        @tmpa = split(' ', $_);
        print DREPORT "$tmpa[15],$tmpa[17],$tmpa[20],$tmp[23],$tmp[26]
+\n";
        $cnt++;
        }

close TMP1;
unlink ("tmpfile.txt");

[download]

Everything works with the exception of bytes xmt/rcv, it seems the data doesn't show up for some reason. Here's an example of the output I get:


[user1],[group1],0:09:19,,


[user2],[group2],0:27:21,,


[user3],[group3],7:07:09,,

Thanks for your time/help.

Edited 2004-08-05 by Ovid

Replies are listed 'Best First'.
Re: Can't get data out of array
by Tomte (Priest) on Aug 05, 2004 at 16:27 UTC

    Always use strict; and use warnings; if you are developing, helps a lot:

    This a little bit "strictified" version of your code 

    #!/bin/env perl
#Look in daily syslog file and grab the
#user, group, connection duration, bytes xmt and bytes rcv
#Stick this data into a daily report in .csv format
#
use warnings;
use strict;

#TIME & DATE STUFF
my $stime = localtime;
my @time1 = split(' ', $stime);
my $mon = $time1[1];
my $day = $time1[2];
my $year = $time1[4];

my $filename = "dailyreport-$mon-$day-$year.csv";

#Open the log file for reading

open (LOG, "</var/log/ciscon.log") || die "Couldn't open log file $!";
open (TMP, ">>tmpfile.txt") || die "Couldn't open tmp file $!";
open (DREPORT, ">>/root/reports/$filename") || die "Couldn't open csv 
+$!";
print DREPORT "DAILY LOG FOR $mon $day $year\n";
print DREPORT "\n";
print DREPORT "USERNAME,GROUP,DURATION,BYTES XMT,BYTES RCV,\n";
print DREPORT "\n";

#Pump data into array and parse for usefull data
#Usefull data goes into tmpfile

while( <LOG>)
        {
        my @tmp = grep {/disconnected/} $_;
        print TMP "@tmp";
        }

#Close stuff we dont need

close TMP;
close LOG;

#Open temp file and stuff into array by spaces print data we
#want into the daily csv file

open (TMP1, "<tmpfile.txt")  || die "Can't read from tmpfile $!";
while( <TMP1> )
        {
        chomp $_;
        my @tmpa = split(' ', $_);
        print DREPORT "$tmpa[15],$tmpa[17],$tmpa[20],$tmp[23],$tmp[26]
+\n";
        }
close TMP1;
unlink ("tmpfile.txt");

    [download]
    yields: 
    [1823]tom@margo perl $ perl -c test3.pl
Global symbol "@tmp" requires explicit package name at test3.pl line 5
+0.
Global symbol "@tmp" requires explicit package name at test3.pl line 5
+0.
test3.pl had compilation errors.

    [download]
    I guess you mean tmpa in both cases, but I may be wrong...

    regards,
    tomte

    An intellectual is someone whose mind watches itself.
    -- Albert Camus

[reply]
[d/l]
[select]
      Ok, I'm a moron for missing that. You can bet I'll have use strict and warnings from now on. Sheesh. Thanks!
[reply]

Back to Seekers of Perl Wisdom