My problem is that I just can't see how to code that structure up.

If I understand what you are trying to do one way I would approach this would be:

my $files = {
   filename1 => {
                   field1 => { start_position => 10,
                               length => 20 
                              }, # values as appropriate
                   field2 => { 
                               start_position => 20
                               length => 5
                              }
                   | 
                   etc. 
                   | 
           },
     filename2 => {
              | and so on...
     }
}
[download]

   foreach my $file ( keys %$files ){
        foreach my $field ( keys %{$files->{$file}}{
        |
        do something here 
        |
[download]

My preferred approach would be object oriented in nature where I'd create a module similar to this:

package MyDataFile;

sub new {
    my $p = shift;
    my $class="MyDataFile";
    my $self = {
          fields => { },
          filename => undef
   };
   bless $self, $class;
   return $self;
}

sub filename {
   my $self = shift;
   if ( $_[0] ) {
       $self->{filename} = shift;
   }
   return $self->{filename};
}

sub addField {
   my ($self,$field_name,$start,$length)=@_;
   $self->{fields} = {
              fieldname=>$field_name,
              start=> $start,
              length=> $length
      };
}

sub getField {
     my ($self,$fieldname)=@_;
     return $self->{fields}->{$fieldname};
}

sub getAllFieldnames {
     my $self=shift;
     return keys %{$self->{fields}};
}


1;
[download]

There are other methods I might add, and I might be tempted as well to create a module defining a structure descriping the field attributes as well, but that's way too complicated for the sake of this discussion for now.

Given the module I just gave you, you could do something like this in your code now:

use strict;
use MyDataFile;

my @files=();

   | 
    mumble...
   |
my $df = new MyDataFile;
$df->filename('foo.blah');
$df->addField('snarf',5,20);

   | 
   etc.
    |
[download]

Hope this is food for thought.

That's exactly what I was looking for :) Thanks!!!

I did consider going OO, but given that I've only got limited time to get this together, I'm going to stick with the first part of your solution.

Thanks, Simon

I think, for the moment you need to step back and design how you are going to use this data.

Figure out how you are going to work with it first. If force the data into some sort of structure without understanding how you are going to use, you will just make problems for yourself.

So to start, give us a sample of code that shows HOW you intend to use the data. Then will be easier for people to show you how to manipulate your data to match your interface.

Something like YAML seems well-suited for your means. The configuration in a YAML file will translate directly to your datastructure.

Plan B was going to be to go with an XML config file, but given that once this is setup it won't get changed very often.

But thanks for the suggestion.

Simon

Plan B was going to be to go with an XML config file...

You mean something like this?

$ cat test.xml
<?xml version="1.0"?>
<filelist>
<file name="file1">
  <field name="field1" start="50"  length="20"/>
  <field name="field2" start="80"  length="20"/>
  <field name="field3" start="100"  length="20"/>
</file>
<file name="file2">
  <field name="field1" start="52"  length="22"/>
  <field name="field2" start="82"  length="22"/>
  <field name="field3" start="120"  length="22"/>
</file>
<file name="file3">
  <field name="field1" start="53"  length="23"/>
  <field name="field2" start="83"  length="23"/>
  <field name="field3" start="130"  length="23"/>
</file>
</filelist>


$ perl -MXML::Simple -MData::Dumper -e '$x=XMLin("test.xml"); print Du
+mper($x)'

$VAR1 = {
          'file' => {
                    'file2' => {
                               'field' => {
                                          'field1' => {
                                                      'length' => '22'
+,
                                                      'start' => '52'
                                                    },
                                          'field2' => {
                                                      'length' => '22'
+,
                                                      'start' => '82'
                                                    },
                                          'field3' => {
                                                      'length' => '22'
+,
                                                      'start' => '120'
                                                    }
                                        }
                             },
                    'file1' => {
                               'field' => {
                                          'field1' => {
                                                      'length' => '20'
+,
                                                      'start' => '50'
                                                    },
                                          'field2' => {
                                                      'length' => '20'
+,
                                                      'start' => '80'
                                                    },
                                          'field3' => {
                                                      'length' => '20'
+,
                                                      'start' => '100'
                                                    }
                                        }
                             },
                    'file3' => {
                               'field' => {
                                          'field1' => {
                                                      'length' => '23'
+,
                                                      'start' => '53'
                                                    },
                                          'field2' => {
                                                      'length' => '23'
+,
                                                      'start' => '83'
                                                    },
                                          'field3' => {
                                                      'length' => '23'
+,
                                                      'start' => '130'
                                                    }
                                        }
                             }
                  }
        };
[download]

But... what? Looks pretty simple to me. I'd rather not have all those data values hard-coded in the perl script itself, and reading them from a basic xml file seems like a no-brainer.

I guess it's a little goofy that there are two layers in XML::Simple's "default" hash structure that are sort of useless ("file" and "field"), but if you tried a different way of structuring the xml (and/or spent some time with the XML::Simple man page), you could probably improve on that. (Or you could just use the default as-is and get done quicker.)