=head1 THE DBIx::Class WAY
Here are a few simple tips that will help you get your bearings with
-DBIx::Class.
+DBIx::Class.
-=head2 Tables become ResultSources
+=head2 Tables become Result classes
-DBIx::Class needs to know what your Table structure looks like. You do that by
-defining L<DBIx::Class::ResultSource>s. Each table gets a ResultSource, which
-defines the Columns it has, along with any Relationships it has to other tables.
-(And oh, so much more besides) The important thing to understand:
+DBIx::Class needs to know what your Table structure looks like. You
+do that by defining Result classes. Result classes are defined by
+calling methods proxied to L<DBIx::Class::ResultSource>. Each Result
+class defines one Table, which defines the Columns it has, along with
+any Relationships it has to other tables. (And oh, so much more
+besides) The important thing to understand:
- A ResultSource == Table
+ A Result class == Table
(most of the time, but just bear with my simplification)
=head2 It's all about the ResultSet
So, we've got some ResultSources defined. Now, we want to actually use those
-definitions to help us translate the queries we need into handy perl objects!
+definitions to help us translate the queries we need into handy perl objects!
Let's say we defined a ResultSource for an "album" table with three columns:
"albumid", "artist", and "title". Any time we want to query this table, we'll
SELECT albumid, artist, title FROM album;
Would be retrieved by creating a ResultSet object from the album table's
-ResultSource, likely by using the "search" method.
+ResultSource, likely by using the "search" method.
DBIx::Class doesn't limit you to creating only simple ResultSets -- if you
wanted to do something like:
SELECT title FROM album GROUP BY title;
-You could easily achieve it.
+You could easily achieve it.
-The important thing to understand:
+The important thing to understand:
- Any time you would reach for a SQL query in DBI, you are
+ Any time you would reach for a SQL query in DBI, you are
creating a DBIx::Class::ResultSet.
=head2 Search is like "prepare"
Setting up a ResultSet does not execute the query; retrieving
the data does.
+=head2 Search results are returned as Rows
+
+Rows of the search from the database are blessed into
+L<DBIx::Class::Row> objects.
+
=head1 SETTING UP DBIx::Class
Let's look at how you can set and use your first native L<DBIx::Class> tree.
use base qw/DBIx::Class::Schema/;
In this class you load your result_source ("table", "model") classes, which we
-will define later, using the load_classes() method. You can specify which
-classes to load manually:
-
- # load My::Schema::Album and My::Schema::Artist
- __PACKAGE__->load_classes(qw/ Album Artist /);
-
-Or load classes by namespace:
+will define later, using the load_namespaces() method:
- # load My::Schema::Album, My::Schema::Artist and My::OtherSchema::LinerNotes
- __PACKAGE__->load_classes(
- {
- 'My::Schema' => [qw/ Album Artist /],
- 'My::OtherSchema' => [qw/ LinerNotes /]
- }
- );
+ # load My::Schema::Result::* and their resultset classes
+ __PACKAGE__->load_namespaces();
-Or let your schema class load all classes in its namespace automatically:
+By default this loads all the Result (Row) classes in the
+My::Schema::Result:: namespace, and also any resultset classes in the
+My::Schema::ResultSet:: namespace (if missing, the resultsets are
+defaulted to be DBIx::Class::ResultSet objects). You can change the
+result and resultset namespaces by using options to the
+L<DBIx::Class::Schema/load_namespaces> call.
- # load My::Schema::*
- __PACKAGE__->load_classes();
+It is also possible to do the same things manually by calling
+C<load_classes> for the Row classes and defining in those classes any
+required resultset classes.
Next, create each of the classes you want to load as specified above:
- package My::Schema::Album;
- use base qw/DBIx::Class/;
+ package My::Schema::Result::Album;
+ use base qw/DBIx::Class::Core/;
-Load any components required by each class with the load_components() method.
-This should consist of "Core" plus any additional components you want to use.
-For example, if you want serial/auto-incrementing primary keys:
+Load any additional components you may need with the load_components() method,
+and provide component configuration if required. For example, if you want
+automatic row ordering:
- __PACKAGE__->load_components(qw/ PK::Auto Core /);
+ __PACKAGE__->load_components(qw/ Ordered /);
+ __PACKAGE__->position_column('rank');
-C<PK::Auto> is supported for many databases; see L<DBIx::Class::Storage::DBI>
-for more information.
+Ordered will refer to a field called 'position' unless otherwise directed. Here you are defining
+the ordering field to be named 'rank'. (NOTE: Insert errors may occur if you use the Ordered
+component, but have not defined a position column or have a 'position' field in your row.)
Set the table for your class:
Add columns to your class:
- __PACKAGE__->add_columns(qw/ albumid artist title /);
+ __PACKAGE__->add_columns(qw/ albumid artist title rank /);
Each column can also be set up with its own accessor, data_type and other pieces
of information that it may be useful to have -- just pass C<add_columns> a hash:
is_auto_increment => 0,
default_value => '',
},
- title =>
+ title =>
{ data_type => 'varchar',
size => 256,
is_nullable => 0,
is_auto_increment => 0,
default_value => '',
+ },
+ rank =>
+ { data_type => 'integer',
+ size => 16,
+ is_nullable => 0,
+ is_auto_increment => 0,
+ default_value => '',
}
);
DBIx::Class doesn't directly use most of this data yet, but various related
modules such as L<DBIx::Class::WebForm> make use of it. Also it allows you to
create your database tables from your Schema, instead of the other way around.
-See L<SQL::Translator> for details.
+See L<DBIx::Class::Schema/deploy> for details.
See L<DBIx::Class::ResultSource> for more details of the possible column
attributes.
-Accessors are created for each column automatically, so My::Schema::Album will
+Accessors are created for each column automatically, so My::Schema::Result::Album will
have albumid() (or album(), when using the accessor), artist() and title()
methods.
make a predefined accessor for fetching objects that contain this Table's
foreign key:
- __PACKAGE__->has_many('albums', 'My::Schema::Artist', 'album_id');
+ # in My::Schema::Result::Artist
+ __PACKAGE__->has_many('albums', 'My::Schema::Result::Album', 'artist');
See L<DBIx::Class::Relationship> for more information about the various types of
available relationships and how you can design your own.
=head2 Connecting
-To connect to your Schema, you need to provide the connection details. The
-arguments are the same as for L<DBI/connect>:
+To connect to your Schema, you need to provide the connection details or a
+database handle.
+
+=head3 Via connection details
+
+The arguments are the same as for L<DBI/connect>:
my $schema = My::Schema->connect('dbi:SQLite:/home/me/myapp/my.db');
Note that L<DBIx::Class::Schema> does not cache connections for you. If you use
multiple connections, you need to do this manually.
-To execute some sql statements on every connect you can add them as an option in
+To execute some SQL statements on every connect you can add them as an option in
a special fifth argument to connect:
my $another_schema = My::Schema->connect(
{ on_connect_do => \@on_connect_sql_statments }
);
-See L<DBIx::Class::Schema::Storage::DBI/connect_info> for more information about
+See L<DBIx::Class::Storage::DBI/connect_info> for more information about
this and other special C<connect>-time options.
+=head3 Via a database handle
+
+The supplied coderef is expected to return a single connected database handle
+(e.g. a L<DBI> C<$dbh>)
+
+ my $schema = My::Schema->connect (
+ sub { Some::DBH::Factory->connect },
+ \%extra_attrs,
+ );
+
=head2 Basic usage
Once you've defined the basic classes, either manually or using
my $album = $schema->resultset('Album')->find(14);
This will run a C<SELECT> with C<albumid = 14> in the C<WHERE> clause, and
-return an instance of C<My::Schema::Album> that represents this row. Once you
+return an instance of C<My::Schema::Result::Album> that represents this row. Once you
have that row, you can access and update columns:
$album->title('Physical Graffiti');
$album->set_column('title', 'Presence');
$title = $album->get_column('title');
-Just like with L<Class::DBI>, you call C<update> to commit your changes to the
-database:
+Just like with L<Class::DBI>, you call C<update> to save your changes to the
+database (by executing the actual C<UPDATE> statement):
$album->update;
=head2 Adding and removing rows
To create a new record in the database, you can use the C<create> method. It
-returns an instance of C<My::Schema::Album> that can be used to access the data
+returns an instance of C<My::Schema::Result::Album> that can be used to access the data
in the new record:
- my $new_album = $schema->resultset('Album')->create({
+ my $new_album = $schema->resultset('Album')->create({
title => 'Wish You Were Here',
artist => 'Pink Floyd'
});
=head1 NOTES
+=head2 The Significance and Importance of Primary Keys
+
+The concept of a L<primary key|DBIx::Class::ResultSource/set_primary_key> in
+DBIx::Class warrants special discussion. The formal definition (which somewhat
+resembles that of a classic RDBMS) is I<a unique constraint that is least
+likely to change after initial row creation>. However this is where the
+similarity ends. Any time you call a CRUD operation on a row (e.g.
+L<delete|DBIx::Class::Row/delete>,
+L<update|DBIx::Class::Row/update>,
+L<discard_changes|DBIx::Class::Row/discard_changes>,
+etc.) DBIx::Class will use the values of of the
+L<primary key|DBIx::Class::ResultSource/set_primary_key> columns to populate
+the C<WHERE> clause necessary to accomplish the operation. This is why it is
+important to declare a L<primary key|DBIx::Class::ResultSource/set_primary_key>
+on all your result sources B<even if the underlying RDBMS does not have one>.
+In a pinch one can always declare each row identifiable by all its columns:
+
+ __PACKAGE__->set_primary_keys (__PACKAGE__->columns);
+
+Note that DBIx::Class is smart enough to store a copy of the PK values before
+any row-object changes take place, so even if you change the values of PK
+columns the C<WHERE> clause will remain correct.
+
+If you elect not to declare a C<primary key>, DBIx::Class will behave correctly
+by throwing exceptions on any row operation that relies on unique identifiable
+rows. If you inherited datasets with multiple identical rows in them, you can
+still operate with such sets provided you only utilize
+L<DBIx::Class::ResultSet> CRUD methods:
+L<search|DBIx::Class::ResultSet/search>,
+L<update|DBIx::Class::ResultSet/update>,
+L<delete|DBIx::Class::ResultSet/delete>
+
+For example, the following would not work (assuming C<People> does not have
+a declared PK):
+
+ my $row = $schema->resultset('People')
+ ->search({ last_name => 'Dantes' })
+ ->next;
+ $row->update({ children => 2 }); # <-- exception thrown because $row isn't
+ # necessarily unique
+
+So instead the following should be done:
+
+ $schema->resultset('People')
+ ->search({ last_name => 'Dantes' })
+ ->update({ children => 2 }); # <-- update's ALL Dantes to have children of 2
+
=head2 Problems on RHEL5/CentOS5
-There is a problem with slow performance of certain DBIx::Class operations in
-perl-5.8.8-10 and later on RedHat and related systems, due to a bad backport of
-a "use overload" related bug. The problem is in the Perl binary itself, not in
-DBIx::Class. If your system has this problem, you will see a warning on
-startup, with some options as to what to do about it.
+There used to be an issue with the system perl on Red Hat Enterprise
+Linux 5, some versions of Fedora and derived systems. Further
+information on this can be found in L<DBIx::Class::Manual::Troubleshooting>
=head1 SEE ALSO