The algorithm won't be hard if you can say what you're aiming for. Doesn't have to be a formal statement of requirements, an example is fine as long as it covers the obvious questions.

• Do you want IDs 000000002 and 000000003 in the output?
• What is the Occurance and why is it 1?

I'll go ahead and assume the input log is sorted by timestamp and in fixed-width format.

# see manpage for unpack
my $TEMPLATE = 'A20 @28A9';

# read the input log into a hash
my %last_time;
while (<DATA>) {
    my ($timestamp, $id) = unpack $TEMPLATE;
    $last_time{$id} = $timestamp;
}

# print the output log 
# I have omitted the header

for my $id (sort keys %last_time) {
    print "$id\t$last_time{$id}\n";
}
[download]