Here's my approach to this problem. I'd use a hash to store this information, which will make it easy to sort on and quick to get back out when it's time to print. Here's an example of how'd I'd do what I think you're asking:

$from = "news.dat.txt";
my %h;

open(FILE, "$from") || die "Can't open $from!\n";
while (<FILE>) {
   chomp;
   ($title,$url,$site,$date,$description) = split(/~/,$_);
   
   # if the date has already been seen, tack this entry on to the hash
+ value
    if( $h{$date} ){ 
        $h{$date} .= "|$title~$url~$site~$description";
    }    
    
    # date hasn't been seen, so create a new hash entry
    else
    {
        $h{$date} = "$title~$url~$site~$description";
    }
}

#sort the dates in desc. order
foreach $key (sort {$b cmp $a} keys %h)
{
   #do for each item for that date
    @items = split '\|', $h{$key};
    foreach $item (@items)
    {
        #split back out the values and print the table
        my($title,$url,$site,$description) = split /~/,$item;
        print "<table width=400><tr>";
        print "<td><a href=\"$url\"><font color=blue>$title</font></a>
+<br><font color=black>$key,$site<br>$description</font></td>";
        print "</tr></table><br>";
    }
}
[download]