It seems to me like you should make a hash of arrays mapping dates to their records. So, you could do something like this:

my %records;
while( my $row = $sth->fetchrow_hashref ) { 
    push @{ $records{ $row->{start_date} } }, $row;
}
[download]

Then you can loop through that hash (sorting as necessary) to output the records. Since the outer hash is keyed by the datestamp, we can sort it easily:

foreach my $date( sort keys %records ) { 
    print $date, "\n";

    # then get each record, sorted by the start time
    my @recs = sort {
        $a->{start_time} cmp $b->{start_time}
    } @{ $records{$date} };

    foreach my $rec( @recs ) { 
        print $rec->{id_number}, "\t", $rec->{start_time};
        print "\n";
    }
}
[download]

That's all untested but should get you on the right track.

As a little workout for my newly acquired CGI muscles I put together:

use strict;
use warnings;
use CGI::Pretty;

my %data;

while (<DATA>) {
    chomp;
    
    my ($code, $date, $time) = split /\s+/;
    
    next if ! defined $time;
    
    push @{$data{$date}}, [$code, $time];
}

my $cgi = CGI->new();
my $tableContents;

for (sort keys %data) {
    $tableContents .= $cgi->Tr ($cgi->td ('&nbsp;'));
    $tableContents .= $cgi->Tr ($cgi->td ($_));
    
    for (sort {$a->[1] cmp $b->[1]} @{$data{$_}}) {
        $tableContents .= $cgi->Tr (
                $cgi->td ($_->[0]), $cgi->td ($_->[1])
            ); 
    };
}

#print $cgi->header ();
print $cgi->start_html ();
print $cgi->table ($tableContents);
print $cgi->end_html ();

__DATA__
CA642997     2007-01-29     03:05:00
CA651167     2007-01-29     03:05:00
CA651038     2007-01-29     19:00:00
CA650995     2007-01-30     19:00:00
CA652568     2007-01-31     01:05:00
CA652880     2007-01-31     01:30:00
CA652884     2007-01-31     06:00:00
CA643602     2007-02-01     00:30:00
CA652881     2007-02-01     01:00:00
[download]

which prints output that renders as (having stripped off headers and stuff for posting here):

my $last_date;
for my $row (@rows) {
  if ($row->{start_date} ne $last_date) {
    print "$row->{start_date}\n";
    $last_date = $row->{start_date};
  }
  printf "%8s %8s", $row->{id_number}, $row->{start_time};
}
[download]

geez i do have brain damage.

what you have written i already figured out (thanks, though). but i completely forgot to add a crucial component to this problem.

if the start_time of the following day is earlier that 09:00, i want it printed with the previous day. for example, if this were the output:

id_number
CA456789   2007-01-29   16:00:00
CA456712   2007-01-30   03:00:00
[download]

i would like it to be output like this:

2007-01-29
CA456789   16:00:00
CA456712   03:00:00
[download]

thanks again for you help

Modify tinita's code as follows

if ($row->{start_date} ne $last_date && (split/:/,$row->start_time)[0]
+ > 9) {
[download]

update: Oops. That will output incorrect results if all entries for a day are from before 9:00. Make a hash keyed with unix timestamps of each date, push each entry onto an array for the appropriate day:

use Time::Local;
...
for my $row (@rows) {
    my @l = reverse split /-/, $row->{start_date};
    $l[1]--; # months range is 0..11
    my $key = timelocal(0,0,0,@l);
    my $line = join("\t", $row->{id_number}, $row->{start_time}); # wh
+atever formatting

    if(split/:/,$row->start_time)[0] > 9) {
        push @{$hash{$key-86400}}, $line;
    } else {
        push @{$hash{$key}}, $line;
    }
}
foreach my $key (sort %hash) {
    my @l = (localtime $key)[5,4,3];
    $l[1]++; # month...
    print join('-', @l), "\n";
    print $_,"\n" for @{$hash{$key}};
}
[download]

--shmem

_($_=" "x(1<<5)."?\n".q·/)Oo.  G°\        /
                              /\_¯/(q    /
----------------------------  \__(m.====·.(_("always off the crowd"))."·
");sub _{s./.($e="'Itrs `mnsgdq Gdbj O`qkdq")=~y/"-y/#-z/;$e.e && print}