There are many different ways you could tackle this depending in large part how the configuration/data base file will be maintained. If you are writing a separate application to manage the file then you could use XML::Simple, Data::Dump::Streamer, a real data base using DBI and DBD::SQLite or a whole range of other options.

If your file format is XML or YAML there is a chance that it could be edited manually without breaking things, with YAML likely to be easier to get on with than XML.

Alternatively you could choose one of the plethora of configuration file formats (often Windows .ini style) and pick a module from CPAN to load it for you (there are many).

Or lastly, you could roll your own application specific file format and parse that. Code to do that using Perl is likely to be pretty small if you are even halfway smart about your file format and field separators.

The key question though is: hand edited or managed. Keep in mind that if you allow it to be hand edited, one day someone is going to screw it up.

Thanks for your help guys. In giving more thought to the problem, what I want to do is not hand edit the flat text file, but rather build an admin menu, so I can enter new services when needed. Also I wanted to get rid of the hash at the beginning of my program, which has the services in it and having to update it by hand. Instead I wanted to populate the hash with whatever is in the flat text file Does anybody know how I can do this?

If the number of entries (for whatever definition of entry) is modest then using Data::Dump::Streamer is likely to be as good as anything. If you anticipate a large number of entries then it's probably worth finding out about DBI and the various DBD drivers that allow you to use a "real" data base.

However, for the simple case consider:

use strict;
use warnings;
use Data::Dump::Streamer;


my %services = (
    1 => {
        name => "service1",
        host => { host1 => 1 },
    },

    2 => {
        name => "service2",
        host => { host0 => 2, host5 => 2 },
    },
);

Dump (\%services);

open TEMP, '>', 'temp.txt';
print TEMP Dump (\%services)->Out ();
close TEMP;

my $newServices = do 'temp.txt';

Dump ($newServices);
[download]

Prints:

$HASH1 = {
           1 => {
                  host => { host1 => 1 },
                  name => 'service1'
                },
           2 => {
                  host => {
                            host0 => 2,
                            host5 => 2
                          },
                  name => 'service2'
                }
         };
$HASH1 = {
           1 => {
                  host => { host1 => 1 },
                  name => 'service1'
                },
           2 => {
                  host => {
                            host0 => 2,
                            host5 => 2
                          },
                  name => 'service2'
                }
         };
[download]

---

DWIM is Perl's answer to Gödel

I've just looked into XML::Simple as suggested and I can see how I can achieve what I wanted the program to do. Thanks guys.

#!/usr/bin/perl

use strict;
use warnings;
use Data::Dumper;
$Data::Dumper::Indent = 1;

my  %services = (
    1 => {
        name => "service1",
        host => { host1 => 1 },
    },

    2 => {
        name => "dxuldm",
        host => { host0 => 2, host5 => 2 },
    },
);

open my $fh, '>', 'services.txt'
  or die "can't open to write: $!\n";
print $fh Data::Dumper->Dump([\%services], [qw(services_restored)]);
close $fh;

our $services_restored;
eval {do 'services.txt'};

print Dumper $services_restored;
[download]

$VAR1 = {
  '1' => {
    'name' => 'service1',
    'host' => {
      'host1' => 1
    }
  },
  '2' => {
    'name' => 'dxuldm',
    'host' => {
      'host0' => 2,
      'host5' => 2
    }
  }
};
[download]

This is all really good advice, but at the moment I have been playing around with XML::Simple. The problem that I am having is I cannot generate the data structure that I had before with the hash it keeps on putting things in a nested array. This is what I used to have before using XML::Simple

$VAR1 = '1';
$VAR2 = {
          'host' => {
                      'hosta' => 1
                    },
          'name' => 'servicea'
        };
$VAR3 = '2';
$VAR4 = {
          'host' => {
                      'host0' => 2,
                      'host5' => 2
                    },
          'name' => 'serviceb'
        };
[download]

When using XML::Simple I get this:

$VAR1 = {
          'name' => 'servicea',
          'host' => [
                    {
                      'hosta' => '2'
                    },
                    {
                      'hostb' => '2'
                    }
                  ]
[download]

I am using the following XML with XML::simple.

<?xml version='1.0'?>
<services>
<name>servicea</name>
<host>
        <hosta>2</hosta>
</host>
<host>
        <hosta>2</hosta>
</host>
</services>
[download]

Is there any way I can change the XML, so XML::simple only uses hash of hashes instead of hashes and arrays? Thanks

*sighs* I provide you with a suggestion. You say "That's cool, but can you debug what's wrong with this other mechanism?" XML is a very poor solution for this, which is why I suggested DBM::Deep. dbm-deep is designed to solve this specific problem that you're working with - easy persistence of Perl datastructures.

Use the proper solution for the job. XML isn't it.

---

My criteria for good software:

1. Does it work?
2. Can someone else come in, make a change, and be reasonably certain no bugs were introduced?