NAME Data::Mapper - An implementation of Data Mapper Pattern described in PofEAA SYNOPSIS # Your mapper class package My::Mapper; use parent qw(Data::Mapper); # Your data class related to `user` table package My::Mapper::Data::User; use parent qw(Data::Mapper::Data); # Then, use them package main; use Data::Mapper::Adapter::DBI; my $dbh = DBI->connect($dsn, $username, $password, ...); my $adapter = Data::Mapper::Adapter::DBI->new({ driver => $dbh }); # You can pass coderef as a driver factory, instead: my $handler = DBIx::Handler->new(...); my $adapter = Data::Mapper::Adapter::DBI->new({ driver => sub { $handler->dbh } }); my $mapper = My::Mapper->new({ adapter => $adapter }); # Create my $data = $mapper->create(user => { name => 'kentaro', age => 34 }); #=> is a My::Mapper::Data::User object # Retrieve just one item $data = $mapper->find(user => { name => 'kentaro' }); #=> is a My::Mapper::Data::User object $data->param('name'); #=> kentaro $data->param('age'); #=> 34 # Search with some conditions $result = $mapper->search(user => { age => 34 }, { order_by => 'id DESC' }); for my $data (@$result) { $data->param('name'); ... } # Update $data->param(age => 35); my $sth = $mapper->update($data); $sth->rows; #=> 1 # Destroy my $sth = $mapper->delete($data); $sth->rows; #=> 1 WARNING This software is under the heavy development and considered ALPHA quality now. Things might be broken, not all features have been implemented, and APIs will be likely to change. YOU HAVE BEEN WARNED. DESCRIPTION Data::Mapper is an implementation of Data Mapper Pattern described in PofEAA, written by Martin Fowler, and is kind of a ORM, but not limited only to it, that is, this module just relates some data to another; for example, data from a database to Perl's objects. Data::Mapper Convention This module, actually, merely defines a simple convention how to make relations between some data to another, and now has only one adapter implementation: Data::Mapper::Adapter::DBI. Mapper *Mapper* makes relations between data from a datasource, which is typically a database, to Perl's objects, and vice versa, while keeping them independent each other, and the mapper itself. You can use Data::Mapper via your own mapper subclass by inheriting it. *Mapper* provides the methods below: * create( *$name* => *\%values* ) Creates a new data, and returns it as a *Data* object described later. * find( *$name* => *\%conditions* [, *\%options*] ) Searches data according to "\%conditions" and "\%options", and returns the first one as a *Data* object described later. * search( *$name*, *\%conditions* [, *\%options*] ) Searches data according to "\%conditions" and "\%options", and returns the all of them as an ArrayRef which contains each records as a *Data* object described later. * update( *$data* ) Updates $data in the datasource. * delete( *$data* ) Deletes the $data from the datasource. Adapter *Adapter* does CRUD operations against a datasource (database, memcached, external API, etc.). It must implement some methods according to the convention. *Adapter* must implements the methods below: * create( *$name*, *\%values* ) Creates a new data, and returns it as a specific form described later. * find( *$name*, *\%conditions* [, *\%options*] ) Searches data according to "\%conditions" and "\%options", and returns the first one as a specific form described later. * search( *$name*, *\%conditions* [, *\%options*] ) Searches data according to "\%conditions" and "\%options", and returns the all of them as an ArrayRef which contains each records as the specific form same as the one "find()" method returns. * update( *$name*, *\%values* [, *\%conditions*] ) Updates data in a datasource according to "\%values", and "\%conditions". * delete( *$name*, *\%conditions* ) Deletes the data specified by "\%conditions" from a datasource. The return value of "create()", "find()", "search()" must be either a plain HashRef or a Hash-based object. If the object has "as_serializable()", it'll be called before mapping to extract data as a HashRef. You can adapt any data-retrieving module to Data::Model convention if only you implement the methods described above. Data *Data* represents a data model where you can define some business logic. You must notice that *Data* layer has no idea about what *Mapper* and *Adapter* are. It just holds the data passed by *Mapper* *Data* can be either *Data::Mapper::Data*-based object or your own POPO. # Data::Mapper::Data-based class package My::Mapper::Data::User; use parent qw(Data::Mapper::Data); # Or, Hash-based POPO package My::Mapper::Data::User; sub new { my ($class, %args) = @_; bless \%args, $class; } package My::Mapper; use parent qw(Data::Mapper); package main; My::Mapper; my $mapper = My::Mapper->new(...); $mapper->find(user => ...) #=> Now returns data as a My::Mapper::Data::User What data class will be used is determined by "Data::Mapper#data_class" method. In default, data class will be "Your::Mapper::Data::$table" as shown above. You can customize the behaviour by overriding the method. AUTHOR Kentaro Kuribayashi REPOSITORY * GitHub SEE ALSO * Data Mapper Pattern * DBIx::ObjectMapper An existing Perl implementation of the pattern above. You might want to consult it if you want much more ORM-ish features. * DBI * SQL::Maker LICENSE Copyright (C) Kentaro Kuribayashi This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.