I'm relatively new to Perl and am having problems creating a hash of arrays. The file, about 500MB in size, is formatted like so:

2011-04-03 09:37:12.129 (INFO, ICELineHandler.cpp:339) Product Def Mar
+ketID 90120253, Symbol 'BRN FMU0012_OMCA0000118502081312'
<b>2011-04-09 21:32:15.525 [3509,3523]: Gap detect on 233.156.208.41:2
+0041 from 2746318 to 2746373, moving to next message
2011-04-09 21:32:15.585 [3509,3523]: Gap detect on 233.156.208.41:2004
+1 from 2746420 to 2746475, moving to next message
2011-04-09 21:32:15.639 [3509,3522]: Received data on Connection[ICE-O
+ptions]. Pending=214044. </b>
2011-04-03 09:37:12.129 (INFO, ICELineHandler.cpp:339) Product Def Mar
+ketID 90120253, Symbol 'BRN FMU0012_OMCA0000118502081312'
[download]

Using the timestamp in the second column as the key, I need to search for any line containing a "Pending=" queue greater than a user specified amount, and add it to an array. Also, if there was a "Gap" detected at the same second (I'll be stripping the milliseconds), add them to the array for that same time stamp. In the text above, each key/array pair to read like:

my %ICE = (
    21:31:10 => [ 'Pending=3201', ],
    21:32:14 => [ 'Pending=1000', ],
    21:32:15 => [ 'Gap detect', '233.156.208.41:20041', 'Pending=21404
+4',],
    21:32:24 => [ 'Gap detect', '233.156.208.41:20041', 'Pending=10400
+0',],
    21:32:58 => [ 'Gap detect', '233.156.208.41:20041' 'Pending=96000'
+,],
    21:33:12 => [ 'Pending=528', ]
);

And print comma separated:

21:31:10, Gap detect 233.156.208.41:20041, Pending=
21:31:12, Gap detect 233.156.208.41:20041, Pending=3400
[download]

Thus far, I have the following code, but it doesn't seem to work. Any recommendations?

UPDATE:
Apologies for the "half-arsed" description and lack of presentation of the results I'm getting. I wrote this in a hurry, thinking I could update it when I got home. Being new here, I didn't expect so many quick responses. I'll update the post shortly. Thanks for your help

# Get today's date which will be used as the default.
my($day, $month, $year) = (localtime)[3,4,5];
$month = sprintf '%02d', $month+1;
$day   = sprintf '%02d', $day;
$year = $year+1900;
$ymd = "$year-$month-$day";

my $total; # variable to be used for queue total for the given timefra
+me
my $count = 0;

## Get command line arguments, convert time to seconds
##my $logFile = "pmmd-ltc-fsrlabs21-mdrc-server-cta.log";
my $logFile = "pmmd-ltc-fsrlabs41-mdrc-server-ice.log";
my $sTime= $ARGV[0];
my @sTime=split(/:/,$sTime); # split start time
my $sSecs=$sTime[0] * 3600 + $sTime[1] * 60 + $sTime[2]; # convert sta
+rt-time to seconds
my $eTime = $ARGV[1];
my @eTime=split(/:/,$eTime); # split end time
my $eSecs=$eTime[0] * 3600 + $eTime[1] * 60 + $eTime[2]; # convert sto
+p-time to seconds
my $tHold = $ARGV[2];

open(LOG, "$logFile") or die "Couldn't open file for processing: $!";

while ( $line = <LOG> ) {
unless (($data[10] =~ m/Pending=/) || ($data[6] =~ m/Gap/)) { next; } 
+# skip elements we don't want
     if ($line =~ m/Pending=/) {
        my @data=split(/ /,$line); # split the line up
        $data[10] =~ s/[A-Za-z=.]//g; # delete "Pending", "=" and "."
        $data[1] =~ s/\..*//g; # delete the millisecond element of the
+ (line)lTime var
        my @lTime=split(/:/,$data[1]); # split the line time
        my $lSecs=$lTime[0] * 3600 + $lTime[1] * 60 + $lTime[2]; # con
+vert line-time to seconds

        if (($data[0] eq $ymd) && ($data[10] >= $tHold) && ($lSecs >= 
+$sSecs) && ($lSecs <= $eSecs))
            {
                   $line = "$data[1],$data[10]";
                }
   }

      elsif ($line =~ m/Gap detect/) {
        my @data=split(/ /,$line); # split the line up
        $data[1] =~ s/\..*//g; # delete the millisecond element of the
+ (line)lTime var
        my @lTime=split(/:/,$data[1]); # split the line time
        my $lSecs=$lTime[0] * 3600 + $lTime[1] * 60 + $lTime[2]; # con
+vert line-time to seconds
                if (($data[0] eq $ymd) && ($lSecs >= $sSecs) && ($lSec
+s <= $eSecs))
                {
                        $line = "$data[1],$data[9]";
                }
   } else { next; }


   ($time, $rest) = split ',', $line, 2;
   #$time =~ s/\..*//g; # delete the millisecond element
   @fields = split ',', $rest;
   $HoA{$time} = [ @fields ];
}

for $time (sort (keys (%HoA)) ) {
   print "$time: @{ $HoA{$time} }\n";
}

close LOG;
[download]