Persistent::Hash - Hash persistence API (0.1)


        use Persistent::Hash;

        my $phash = Persistent::Hash->new();

        $phash->{key} = 'value';
        $phash->{indexkey} = 'value';
        my $hash_id = $phash->Save();

        my $reloaded_phash = Persistent::Hash->load($hash_id);




This file only gives an overview of Persistent::Hash, to fully grasp it's potential, please consult the Persistent::Hash::API (API Reference) and the Persistent::Hash::Manual (Programmer Manual)


Persistent::Hash is a base class that implements a hash persistence framework for Perl. The Persistent::Hash class implements a TIE mechanism to transform a standard hash into a storable object. It also includes standard method to load and save the object. A Persistent::Hash basically has the same behaviour a normal hash would, but has additionnal methods pertaining to persistence.

The base class uses "storage modules" to know how to save a specific object to storage container and how to reload it, so the whole "storage" part of the implementation is completely modular. A Persistent::Hash can be saved in two chunk, data and index. The "data" chunk is saved as a blob while the "index" chunk is saved in a properly indexed table to take advantage of native database implementations.

More importantly, Persistent::Hash works uniquely by overloading and the interface is made to work so that programmers can easely create their subclasses (called data types) simply by creating a module that ISA Persistent::Hash.


The basic implementation of Persistent::Hash uses the perltie mechanism to hook into the standard hash structure and provide additionnal functionnality. When creating a subclass (data type), you basically create a class that inherits from Persistent::Hash. You control the options of your data type by overloading constants/subroutines to the desired behaviour.



Here is a very simple data type:

        package MyProject::Customer;

        use strict;

        use base qw(Persistent::Hash);

        use constant PROJECT => 'MyProject';
        use constant STORABLE => 1;
        use constant STRICT_FIELDS => 0;


This is your basic class definition for a data type. This one is very minimal and doesn't specify anything. You can see that we defined 3 overloaded constants to create this type, PROJECT, STORAGE and STRICT_FIELDS.

PROJECT specify wich project this data type is part of, this is used to create a MIME type style "type name" for this object that will be used to recognize what type of hash this is when it is retrieved from storage. The hash type for this class for example would be "MyProject/MyProject_Customer". The hash type can be retrieved for an object by using the Type() method.

The STORABLE constants is used to specify if this hash should be allowed to be stored. Naturally we turn this on ;) The STRICT_FIELDS constant is used to control wether Persistent::Hash should control wich keys are set in the hash from a list of allowed keys. (This will be discussed later). In this example, we don't want any strict field attribues.

This object can then be instantiated with the new() constructor:


        use strict;
        use MyProject::Customer;

        my $customer = MyProject::Customer->new();

        $customer->{name} = 'OReilly and Associates';
        $customer->{website} = '';
        $cusomter->{email} = '';

Once you have an instance of your hash, you can use it as you would a standard hash. Set keys in, delete keys, sort the keys...anything. Then, you can save your object using the Save() method.

        my $customer_id = $customer->Save();

The Save() method returns the generated "id" that the storage module returned (typically, if saving to an SQL database, the 'id' will be the next available id in the sequence). Your application can take this id and save it in session, or reference it in another object for later retrieval. You can reload the hash from your storage by using the Load() constructor with the id:

        my $reloaded_customer = MyProject::Customer->Load($customer_id);

The reloaded object is exactly the same as the one you saved (it is a different "perl object", but it contains the same data). You can modify this new hash, and call Save() again, it will update the same record. Now, this customer can be deleted from storage using the Delete() method:


At this point, the has is unusable, it's been untied and the records in storage have been deleted.

This hash was just a very simple example to show you how the actual API looks like, the hash in the storage container (aka database) actually is only a serialized version of the perl reference, not very useful, especially if you want to have performance when retrieving this object from storage.


Let's take our example in the previous chapter and try to make it a little more useful by adding index and strict fields properties:

        package MyProject::Customer;

        use strict;
        use base qw(Persistent::Hash);

        use constant STORABLE => 1;
        use constant PROJECT => 'MyProject';
        use constant STRICT_FIELDS => 1;

        use constant DATA_FIELDS => ['address','phone','comments'];

        use constant INDEX_TABLE => 'myproject_customer_index';
        use constant INDEX_FIELDS => ['name','website','email'];


This object is a little more complex than the previous one because we want to be able run queries on some fields to allow us some very fast searches. In this example, we used STRICT_FIELDS 'on' to make sure that our hash will not get polluted by bad key settings, so only keys listed in the DATA_FIELDS and INDEX_FIELDS will be allowed in the hash.

DATA_FIELDS here is used to define wich fields will be saved in the serialized view of the hash in the datbase. Usually, DATA_FIELDS are fields that you don't want to run search on, because, for example if you are saving to MySQL, searching blobs is not very efficient ;)

INDEX_TABLE specifies the table in wich the index fields will be saved. You want this table to have all the fields listed in INDEX_FIELDS as columns and that table must have an 'id' field to link with the actual hash object.

Now that we have our data type defined, what will happen when saving it ?

Well, the keys listed in DATA_FIELDS will be serialized and saved, the INDEX_FIELDS will be inserted in their respective columns in INDEX_TABLE. When retrieving the object using Load(), those two chunks of data will be reconstructed as one single hash automagically.


This file only contains an overview of Persistent::Hash. The complete API Reference and programmer manual are included in this package in Persistent::Hash::API and Persistent::Hash::Manual. These documents contain more examples and definitions of all the possible configuration options and their explanations.

You can also consult the Storage module programmer's guide: Persistent::Hash::Storage

Please refer to these when developping application and systems using Persistent::Hash.

Online versions of these manuals are available at:

BUGS (known)

None, for now.


The latest developments and changes history for this module are available through cvsweb at:

The bleeding edge code is also available through anonymous CVS access via:

cvs checkout Persistent-Hash

Please, if you use this module in a project, let me know!



Copyright(c) 2001 Benoit Beausejour <>

All rights reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.


Benoit Beausejour <>


Your name here!


perltie(1). perl(1).