Well, an easy way to do it would be to map the month to a quarter with an array:

my @quarter = qw(q0
   q3 q3
   q4 q4 q4
   q1 q1 q1 
   q2 q2 q2
   q3);
# Later...
@sorted = sort
    { $quarter[(split(/\//, $a))[0]] cmp
      $quarter[(split(/\//, $b))[0]] }
        @rawdates;
[download]

Note: I'm assuming all the dates are in the same year, although it shouldn't be hard to extend this to sort on year also.

HTH

I think the following would work. I didn't try to adjust for null dates as I have no idea what your data format is. The snippet sorts by year, then by quarter, then by day.

use strict;
use warnings;
use Data::Dumper;

my @sorted = map  { [ $_->[0], join( '/', @{$_}[1..3] ), $_->[4] ] }
             sort { $a->[3] <=> $b->[3] || $a->[0] <=> $b->[0] || $a->
+[2] <=> $b->[2] }
             map  { [ get_quarter( $_->[0] ), split( '/', $_->[0] ), $
+_->[1] ] }
             map  { [ split( /\s/, $_, 2 ) ] } <DATA>;

chomp @$_ foreach @sorted;

print Dumper \@sorted;

sub get_quarter  {
    my @quarters = qw( 3 3 4 4 4 1 1 1 2 2 2 3 );
    my ( $month ) = (split'/',$_[0])[0];
    $quarters[ $month - 1 ];
}

__DATA__
10/23/2000 data1 foo bar baz
11/19/2000 data2 some stuff
08/11/2001 data3 what the heck?
04/30/2001 data4 this is getting dull
01/01/2001 data5 ??
[download]

Here's what it does:

• It splits the data on whitespace. That leaves the date in the first field. Note the optional third parameter to split which ensures that all data after the date is left untouched.
• It passes the first field to the sub get_quarter() and returns the quarter, along with the date split on the delimeter and the remaining data.
• It sorts on year, then quarter, then day.
• It puts the data back together.

Each array ref is going to have the quarter as the first element, the date as the second and the data as the third.

You're going to have to play with the above snippet as you need to account for null dates. Further, I really didn't test this for multiple quarters spanning multiple years (I assumed you'd have one fiscal year of data at a time).

Cheers,
Ovid

Vote for paco!

Join the Perlmonks Setiathome Group or just click on the the link and check out our stats.

#!/usr/bin/perl
use strict;

# reports to sort - you will need to parse existing to build HoH
my %report;

# dummy data
$report{'report_1'}{'date'} = '10/29/2001';
$report{'report_1'}{'content'} = 'whatever';

$report{'report_2'}{'date'} = '12/29/2001';
$report{'report_2'}{'content'} = 'whatever2';

$report{'report_3'}{'date'} = '4/29/2001';
$report{'report_3'}{'content'} = 'whatever3';

$report{'report_4'}{'date'} = '';
$report{'report_4'}{'content'} = 'whatever4';

# define quarters
my %quarter = (1   => 'q3',  2   => 'q3',   3  => 'q4',
               4   => 'q4',  5   => 'q4',   6  => 'q1',
               7   => 'q1',  8   => 'q1',   9  => 'q2',
               10  => 'q2', 11   => 'q2',  12  => 'q3');
               
my %sorted_reports;
for (keys %report) {
  # grab month
  if ($report{$_}{'date'} =~ m|^(\d+)|) {
    # work out quarter
    # force number - covers 01/1 inconsistancy
    my $q = $quarter{$1+0};

    # whack into hash of arrays
    push @{$sorted_reports{$q}}, $report{$_}{'content'};
  }
  else {
    # no date
    push @{$sorted_reports{'undated reports'}}, $report{$_}{'content'}
+;
  }
}

# now print (or whatever)
# note can sort because 'u' follows 'q'
foreach my $quarter (sort { $a cmp $b; } keys %sorted_reports) {
  # print quarter
  print "\n\n$quarter:\n\n";
 
  # and reports
  print join "\n\n----------\n\n", @{$sorted_reports{$quarter}};

  print "\n\n", '=' x 60;
}
[download]

$_.=($=+(6<<1));print(chr(my$a=$_));$^H=$_+$_;$_=$^H;
print chr($_-39); # Easy but its ok.
[download]

I wonder if you may be reaching for the wrong hammer. Unless you have some requirements that aren't stated, the problem you face is how to find "breaks" in a sorted list of calendar entries, where the breaks correspond to changes in quarter.

Sorting by date is easy. There are some pointers in this thread that'll get you there.

Given the date format you're using (month/day/year), spotting breaks in a sorted list is also easy: Process the list in order. For each entry, extract the month and use it as a key to lookup the quarter in an array or hash, and do your break whenever the quarter changes. The basics of that technique are touched on in Vsarkiss's and CLive ;-)'s posts above.

The code below is less elegant than others in this thread. I understand qw// better now...

However, I do attempt to use the supplied sample data, account for 1 or 2 digit months and days and puke if I don't get a 4 digit year.

#!/usr/bin/perl -w

use strict;

my @test=(    "example: 10/29/2001",
        "example: 2/18/2000 3rd quarter",
        "example: 3/18/2000 4th quarter",
        "example: 4/18/2000 4th quarter",
        "example: 5/18/2000 4th quarter",
        "example: 6/18/2000 1st quarter",
        "example: 7/18/2000 1st quarter",
        "example: 11/18/2001 2nd quarter",
        "example: 01/18/1999 3rd quarter",
        "example: 02/18/1999 3rd quarter",
        "example: 3/18/1999 4th quarter",
        "example: 4/18/1999 4th quarter",
        "example: 5/18/1999 4th quarter",
        "example: 6/18/1999 1 quarter",
        "example: 7/18/1999 1 quarter",
        "example: 8/18/1999 1 quarter",
        "example: 9/18/1999 2nd quarter",
        "example: 10/18/1999 2nd quarter",
        "example: 11/18/1999 2nd quarter",
        "example: 12/18/1999 3rd quarter"
    );

sub my_compare(){
    # use "our" to limit scope to function but persist
    # map months to quarters
    our %quarters=(
        '6'=> 1,
        '06'=> 1,
        '7'=> 1,
        '07'=> 1,
        '8'=> 1,
        '08'=> 1,
        '9'=> 2,
        '09'=> 2,
        '10'=> 2,
        '11'=> 2,
        '12'=> 3,
        '1'=> 3,
        '01'=> 3,
        '2'=> 3,
        '02'=> 3,
        '3'=> 4,
        '03'=> 4,
        '4'=> 4,
        '04'=> 4,
        '5'=> 4,
        '05'=> 4
        );

# $a  and $b are given to us by sort
# they are the left & right sides of the comparision
# 
    my ($mon, $day);

    $a=~m|(\d*)/(\d*)/(\d*)|;
    $mon=length($1)<2?'0'.$1:$1;
    $day=length($2)<2?'0'.$2:$2;
    my $a3=$3 . $quarters{$mon}.$mon .$day;        
    die "bad year in $a" unless length ($3)==4;

    $b=~m|(\d*)/(\d*)/(\d*)|;
    $mon=length($1)<2?'0'.$1:$1;
    $day=length($2)<2?'0'.$2:$2;
    my $b3=$3 . $quarters{$mon}. $mon .$day;
    die "bad year in $b" unless length ($3)==4;

    return $a3 <=> $b3;
}

@test=sort(my_compare(), @test);
print join "\n", @test;
[download]