Connecting Tech Pros Worldwide Forums | Help | Site Map
bytes > topic > perl > answers

Hash sort weirdness

Newbie
  
Join Date: Apr 2008
Posts: 2
#1: Apr 1 '08
I am having trouble sorting secondary keys in a multi-level hash. I have a sample program that exhibits the problem...the following code is sorting New 6 Month before New 1 Month. What am I doing wrong?

Thanks in advance.

Expand|Select|Wrap|Line Numbers
  1. #!/usr/bin/perl
  2.  
  3.  
  4. my %data = (
  5.        "2008-03-24" => {
  6.                  "New 1 Month Subscription" => {
  7.                      id => 1,
  8.                      value => 10, 
  9.                  },
  10.                  "New 12 Month Subscription" => {
  11.                      id => 5,
  12.                      value => 12,
  13.                  },
  14.                  "New 14 Day Trial Subscription" => {
  15.                      id => 3,
  16.                      value => 15,
  17.                  },
  18.                  "New 6 Month Subscription" => {
  19.                      id => 6,
  20.                      value => 20,
  21.                  },
  22.        },
  23.        "2008-03-25" => {
  24.                  "New 1 Month Subscription" => {
  25.                      id => 1,
  26.                      value => 10, 
  27.                  },
  28.                  "New 12 Month Subscription" => {
  29.                      id => 5,
  30.                      value => 12,
  31.                  },
  32.                  "New 3 Month Subscription" => {
  33.                      id => 7,
  34.                      value => 8,
  35.                  },
  36.        },
  37. );
  38.  
  39. foreach my $this_date (sort keys %data) {
  40.   print "$this_date\n";
  41.   foreach my $name ( sort {lc($data{$a}) cmp lc($data{$b})} keys %{ $data{$this_date} } ) {
  42.      print "$name\t";
  43.      print "$data{$this_date}{$name}{id}\n";
  44.   }
  45. }
  46.  

KevinADC's Avatar
Expert
  
Join Date: Jan 2007
Location: Southern California USA
Posts: 4,091
#2: Apr 1 '08

re: Hash sort weirdness

This line is wrong:

Expand|Select|Wrap|Line Numbers
  1. foreach my $name ( sort {lc($data{$a}) cmp lc($data{$b})} keys %{ 
should be:

Expand|Select|Wrap|Line Numbers
  1.   foreach my $name ( sort { $data{$this_date}{$a} cmp $data{$this_date}{$b} } keys %{ $data{$this_date} } ) {
but don't expect 6 to come after 12 when you sort using "cmp" because the sort is what is called asciibetical (sort of like alphabetical), in which case 12 comes before 6 like 'AC' comes before 'D'.

Note: there is no need to use lc() in your sort because all the characters are the same mix of upper and lower case characters. Use lc() or uc() when the strings being sorted are not all the same case.
Newbie
  
Join Date: Apr 2008
Posts: 2
#3: Apr 1 '08

re: Hash sort weirdness

Ahh...I missed that. Now that I have that part corrected...I have tried to sort using both cmp and <=>...both are returning the same results which is:

2008-03-24
New 1 Month Subscription 1
New 12 Month Subscription 5
New 14 Day Trial Subscription 3
New 6 Month Subscription 6

I understand that is it using the ASCII value to sort right now, which is not what I want. How can I get it to sort the string part and then the numbers as numbers?

Thanks again...

Quote:

Originally Posted by KevinADC

This line is wrong:

Expand|Select|Wrap|Line Numbers
  1. foreach my $name ( sort {lc($data{$a}) cmp lc($data{$b})} keys %{ 
should be:

Expand|Select|Wrap|Line Numbers
  1.   foreach my $name ( sort { $data{$this_date}{$a} cmp $data{$this_date}{$b} } keys %{ $data{$this_date} } ) {
but don't expect 6 to come after 12 when you sort using "cmp" because the sort is what is called asciibetical (sort of like alphabetical), in which case 12 comes before 6 like 'AC' comes before 'D'.

Note: there is no need to use lc() in your sort because all the characters are the same mix of upper and lower case characters. Use lc() or uc() when the strings being sorted are not all the same case.
Ganon11's Avatar
Moderator
  
Join Date: Oct 2006
Location: New York, United States of America
Posts: 3,428
#4: Apr 1 '08

re: Hash sort weirdness

You can write your own short subroutine to sort them. In the subroutine, $a will be the first element and $b will be the second element. You take $a and $b and parse them for the number (if you know that's the only place they will differ), and then return the comparison of those numbers.

The easiest way to parse the strings will be a regex - just match for the proper pattern and capture each number in $1.
KevinADC's Avatar
Expert
  
Join Date: Jan 2007
Location: Southern California USA
Posts: 4,091
#5: Apr 1 '08

re: Hash sort weirdness

Quote:

Originally Posted by cici

Ahh...I missed that. Now that I have that part corrected...I have tried to sort using both cmp and <=>...both are returning the same results which is:

2008-03-24
New 1 Month Subscription 1
New 12 Month Subscription 5
New 14 Day Trial Subscription 3
New 6 Month Subscription 6

I understand that is it using the ASCII value to sort right now, which is not what I want. How can I get it to sort the string part and then the numbers as numbers?

Thanks again...


This is a job for the Schwartzian Transform
Reply

« previous question | next question »

/bytes/development

/bytes/it

Forums
Visit our community forums for general discussions and latest on Bytes

Our staff and community areas have moved to bytes.com/forums/

/bytes/about

We are a network of experts and professionals in IT and software development that help one another with answers to tough questions and share insights. Get the best answers to your questions from over 226,449 network members.

dev questions:
algorithms | asp | asp.net | c/c++ | coldfusion | c# | flash | html/css | java | javascript |
mobile dev | .net | perl | php | python | ruby | software dev | vb.net | visual basic | xml
it questions:
access | apache | careers | db2 | iis | macosx | mysql | networking | oracle | postgresql | sql server | unix | windows

About Bytes | Copyright 1995-2009 BYTES. All rights Reserved