Here's a little hack that might get you on track. I had to read the input twice so that I can process it. I'm sure there is better way to do this.

#!/usr/bin/perl

use strict;

my $data_pos = tell DATA;
my %record_of;

while (my $line = <DATA>) {
    chomp( $line );
    my ($key, $val) = (split /\s+/, $line)[3,4];
    $record_of{$key}++;
}

seek DATA, $data_pos, 0;

while (my $line = <DATA>) {
    chomp( $line );
    my ($key, $val) = (split /\s+/, $line)[3,4];

    print "$line  one\n" if ( $record_of{$key} == 1 );
    print "$line  end\n" if ( $record_of{$key} == 0 );

    if ( $record_of{$key} == 2 ) {
        $record_of{$key} = 0;
        print "$line start\n";
        next;
    }
}

__DATA__
1    2    -    perlmonks1    sdweq23
3    4    -    javaforums2    sd341
5    6    -    dreamtech3    fr4frf34
7    8    +    dreamtech3    dcvdf3443
9    10    +    sunforum4    sdcds3443
11    12    -    sunforum4    ddfvdv333
[download]