until (eof FILE) {
  for my $count (3, 2) {
    my @set = map scalar <FILE>, 1 .. $count;
    my ($time, @totals);

    for (@set) {
      chomp;
      my @f = split /\t/;
      $time ||= shift @f;  # only want the FIRST time
      $totals[$_] += $f[$_] for 0 .. $#f;
    }

    print OUT join("\t", $time, map $_/$count, @totals), "\n";
  }
}
[download]

_____________________________________________________
Jeff[japhy]Pinyan: Perl, regex, and perl hacker, who'd like a job (NYC-area)
s++=END;++y(;-P)}y js++=;shajsj<++y(p-q)}?print:??;

Accumulate the values into a hash where the keys are the 5 minute slots. Split every 3rd set of values between the 2 affected slots. Divide by 5 for the averages:

#! perl -slw
use strict;
use Data::Dumper;

my $re_line = qr[ ( \d\d ): ( \d\d ) ( .* $ ) ]x;

my %timeslots;
my $counter = 0;
while( <DATA> =~ $re_line ) { # Extract hrs, mins, data into $1, $2, $
+3
    my @temp = split ' ', $3;

    for( 0 .. 4 ) {
        # split every 3rd line between 2 5-minute slots
        if( $counter++ % 3 == 2 ) {
            $timeslots{ int(     ( $1 * 60 + $2 ) / 5 ) }->[ $_ ] += $
+temp[ $_ ] / 2;
            $timeslots{ int( 1 + ( $1 * 60 + $2 ) / 5 ) }->[ $_ ] += $
+temp[ $_ ] / 2;
        }
        else {
            $timeslots{ int(     ( $1 * 60 + $2 ) / 5 ) }->[ $_ ] += $
+temp[ $_ ];
        }
    }
}
for my $ts ( sort{ $a <=> $b } keys %timeslots ) {
    printf '%2d:%02d' . ' [%f]' x 5 . $/,
        ( $ts * 5 ) / 60,
        ( $ts * 5 ) % 60,
        map{ $_ / 5 } @{ $timeslots{ $ts } } ;
}

__DATA__
08/07/03.22:55  1.029   1.172   1.03    0.086   0.382
08/07/03.22:57  0.829   2.284   1.219   0.087   0.439
08/07/03.22:59  2.437   0.792   0.809   0.087   0.305
08/07/03.23:01  0.653   1.089   0.541   0.116   0.351
08/07/03.23:03  0.823   2.407   0.826   0.04    0.23
08/07/03.23:05  0.797   1.016   0.619   0.195   0.274
08/07/03.23:07  1.742   0.901   1.078   0.087   0.328
08/07/03.23:09  0.897   1.218   0.512   0.096   0.252
08/07/03.23:11  1.146   1.281   0.521   0.086   0.276
08/07/03.23:13  0.924   1.129   0.891   0.4     0.456
08/07/03.23:15  1.103   1.383   1.645   0.09    0.387
08/07/03.23:17  0.86    2.078   1.098   0.635   0.36
08/07/03.23:19  3.832   1.911   0.808   0.086   0.309
[download]

Output

P:\test>282224
22:55 [0.776100] [0.770400] [0.508600] [0.043300] [0.194700]
23:00 [0.295800] [0.778400] [0.322300] [0.035900] [0.146700]
23:05 [0.679800] [0.525400] [0.388100] [0.070000] [0.143400]
23:10 [0.503700] [0.455500] [0.301100] [0.106800] [0.146200]
23:15 [1.048700] [0.994700] [0.718500] [0.153200] [0.202800]
23:20 [0.110300] [0.207800] [0.080800] [0.009000] [0.036000]
[download]

You might have a look at GD::Graph to produce a bar graph for you. As for five minute chunks, since you don't have any actual logging every five minutes it seems like you would always be grabbing every other entry, so why not just grab every other entry? I mean you can still say its 5 minutes because you wouldn't have any data to get it any closer.

Hope that helps

open OUT, ">$Output_Directory/sitescope.dat" or die "Can't open OUTPUT
+ file: $!";
foreach (@data) {
        $counter++;
        @tmp = split(/\t/);
        ($time,$date) = split(/ /,$tmp[0]);
        $date =~ s/2003/03/g;
        $time = substr($time,0,5);
        $mday = substr($date,3,2);
        ($hour,$min) = split(/:/,$time);
        $mod = int($min/5);
        if ($mod != $last_mod) {
                print OUT "$last_date\.$last_time\t";
                foreach (1..5) {
                        $seconds = $seconds{$_}/$counter;
                        printf OUT "%1.2f\t",$seconds;
                        $seconds{$_}=0;
                }
                print OUT "\n";
                $counter=0;
        }
        foreach (@tmp[1..5]) {
                $place++;
                $seconds{$place} += ($_/1000);
        }
        $place=0;
        $last_mod = $mod; $last_date = $date; $last_time = $time;
}
close OUT;
[download]

08/08/03.07:55  1.61    1.158   0.71    0.209   0.535   
08/08/03.07:57  1.094   0.887   0.96    0.126   0.322   
08/08/03.07:59  5.131   0.986   1.029   0.095   0.251   
08/08/03.08:01  1.071   1.274   0.638   0.197   0.347   
08/08/03.08:03  1.166   0.984   0.598   0.092   0.583   

Chunked data:

08/08/03.07:59  2.39    0.94    0.80    0.11    0.30    
08/08/03.08:03  2.31    1.60    1.02    0.20    0.61    

Among other problems:

($time,$date) = split(/ /,$tmp[0]);

I don't think this line is doing what you think its doing. The contents of @tmp are something like

(08/08/03.08:01, 1.071, 1.274, 0.638, 0.197, 0.347)

...and you are trying to split the first element in the array, 08/08/03.08:01, on a non-existent space. This means that $time contains the whole string, while $date is empty!

I think you need something like this:

my ( $date, $time ) = split( /\./, @tmp[0] );

And note that I've changed the order of your scalar variables around!

my best advice to you is:

use strict;
use warnings;
[download]

dave

($time,$date) = split(/ /,$tmp[0]);
This is actually coming from the raw data file before any manipulation which looks like this:

10:40:32 08/08/2003     good    website     servername       8.32 sec,
+ 9 steps, 241K total, 35 images        
16:72472        200     8328    ok      1557    662     1275    582   
+  517     1129    1053    962     
591
[download]


I convert the time/date to the new format for use with Ploticus (an opensource graphing tool which I love) which likes to see a period between the two.

Edit by tye, change PRE to CODE around long lines