The London Perl and Raku Workshop takes place on 26th Oct 2024. If your company depends on Perl, please consider sponsoring and/or attending.


DBIx::Class::Schema - composable schemas


  package Library::Schema;
  use base qw/DBIx::Class::Schema/;

  # load all Result classes in Library/Schema/Result/

  package Library::Schema::Result::CD;
  use base qw/DBIx::Class/;
  __PACKAGE__->load_components(qw/Core/); # for example

  # Elsewhere in your code:
  my $schema1 = Library::Schema->connect(
    { AutoCommit => 1 },

  my $schema2 = Library::Schema->connect($coderef_returning_dbh);

  # fetch objects using Library::Schema::Result::DVD
  my $resultset = $schema1->resultset('DVD')->search( ... );
  my @dvd_objects = $schema2->resultset('DVD')->search( ... );


Creates database classes based on a schema. This is the recommended way to use DBIx::Class and allows you to use more than one concurrent connection with your classes.

NB: If you're used to Class::DBI it's worth reading the "SYNOPSIS" carefully, as DBIx::Class does things a little differently. Note in particular which module inherits off which.



Arguments: %options?

   result_namespace => 'Res',
   resultset_namespace => 'RSet',
   default_resultset_class => '+MyDB::Othernamespace::RSet',

With no arguments, this method uses Module::Find to load all your Result classes from a sub-namespace Result under your Schema class' namespace. Eg. With a Schema of MyDB::Schema all files in MyDB::Schema::Result are assumed to be Result classes.

It also finds all ResultSet classes in the namespace ResultSet and loads them into the appropriate Result classes using for you. The matching is done by assuming the package name of the ResultSet class is the same as that of the Result class.

You will be warned if ResultSet classes are discovered for which there are no matching Result classes like this:

  load_namespaces found ResultSet class $classname with no corresponding Result class

If a Result class is found to already have a ResultSet class set using "resultset_class" to some other class, you will be warned like this:

  We found ResultSet class '$rs_class' for '$result', but it seems 
  that you had already set '$result' to use '$rs_set' instead

Both of the sub-namespaces are configurable if you don't like the defaults, via the options result_namespace and resultset_namespace.

If (and only if) you specify the option default_resultset_class, any found Result classes for which we do not find a corresponding ResultSet class will have their resultset_class set to default_resultset_class.

All of the namespace and classname options to this method are relative to the schema classname by default. To specify a fully-qualified name, prefix it with a literal +.


  # load My::Schema::Result::CD, My::Schema::Result::Artist,
  #    My::Schema::ResultSet::CD, etc...

  # Override everything to use ugly names.
  # In this example, if there is a My::Schema::Res::Foo, but no matching
  #   My::Schema::RSets::Foo, then Foo will have its
  #   resultset_class set to My::Schema::RSetBase
    result_namespace => 'Res',
    resultset_namespace => 'RSets',
    default_resultset_class => 'RSetBase',

  # Put things in other namespaces
    result_namespace => '+Some::Place::Results',
    resultset_namespace => '+Another::Place::RSets',

If you'd like to use multiple namespaces of each type, simply use an arrayref of namespaces for that option. In the case that the same result (or resultset) class exists in multiple namespaces, the latter entries in your list of namespaces will override earlier ones.

    # My::Schema::Results_C::Foo takes precedence over My::Schema::Results_B::Foo :
    result_namespace => [ 'Results_A', 'Results_B', 'Results_C' ],
    resultset_namespace => [ '+Some::Place::RSets', 'RSets' ],


Arguments: @classes?, { $namespace => [ @classes ] }+

"load_classes" is an alternative method to "load_namespaces", both of which serve similar purposes, each with different advantages and disadvantages. In the general case you should use "load_namespaces", unless you need to be able to specify that only specific classes are loaded at runtime.

With no arguments, this method uses Module::Find to find all classes under the schema's namespace. Otherwise, this method loads the classes you specify (using use), and registers them (using "register_class").

It is possible to comment out classes with a leading #, but note that perl will think it's a mistake (trying to use a comment in a qw list), so you'll need to add no warnings 'qw'; before your load_classes call.

If any classes found do not appear to be Result class files, you will get the following warning:

   Failed to load $comp_class. Can't find source_name method. Is 
   $comp_class really a full DBIC result class? Fix it, move it elsewhere,
   or make your load_classes call more specific.


  My::Schema->load_classes(); # loads My::Schema::CD, My::Schema::Artist,
                              # etc. (anything under the My::Schema namespace)

  # loads My::Schema::CD, My::Schema::Artist, Other::Namespace::Producer but
  # not Other::Namespace::LinerNotes nor My::Schema::Track
  My::Schema->load_classes(qw/ CD Artist #Track /, {
    Other::Namespace => [qw/ Producer #LinerNotes /],


Arguments: $storage_type|{$storage_type, \%args}
Return value: $storage_type|{$storage_type, \%args}
Default value: DBIx::Class::Storage::DBI

Set the storage class that will be instantiated when "connect" is called. If the classname starts with ::, the prefix DBIx::Class::Storage is assumed by "connect".

You want to use this to set subclasses of DBIx::Class::Storage::DBI in cases where the appropriate subclass is not autodetected, such as when dealing with MSSQL via DBD::Sybase, in which case you'd set it to ::DBI::Sybase::MSSQL.

If your storage type requires instantiation arguments, those are defined as a second argument in the form of a hashref and the entire value needs to be wrapped into an arrayref or a hashref. We support both types of refs here in order to play nice with your Config::[class] or your choice. See DBIx::Class::Storage::DBI::Replicated for an example of this.


Arguments: $code_reference
Return value: $code_reference
Default value: None

If exception_action is set for this class/object, "throw_exception" will prefer to call this code reference with the exception as an argument, rather than "throw" in DBIx::Class::Exception.

Your subroutine should probably just wrap the error in the exception object/class of your choosing and rethrow. If, against all sage advice, you'd like your exception_action to suppress a particular exception completely, simply have it return true.


   package My::Schema;
   use base qw/DBIx::Class::Schema/;
   use My::ExceptionClass;
   __PACKAGE__->exception_action(sub { My::ExceptionClass->throw(@_) });

   # or:
   my $schema_obj = My::Schema->connect( .... );
   $schema_obj->exception_action(sub { My::ExceptionClass->throw(@_) });

   # suppress all exceptions, like a moron:
   $schema_obj->exception_action(sub { 1 });


Arguments: boolean

Whether "throw_exception" should include stack trace information. Defaults to false normally, but defaults to true if $ENV{DBIC_TRACE} is true.


Arguments: $sqlt_schema

An optional sub which you can declare in your own Schema class that will get passed the SQL::Translator::Schema object when you deploy the schema via "create_ddl_dir" or "deploy".

For an example of what you can do with this, see "Adding Indexes And Functions To Your SQL" in DBIx::Class::Manual::Cookbook.

Note that sqlt_deploy_hook is called by "deployment_statements", which in turn is called before "deploy". Therefore the hook can be used only to manipulate the SQL::Translator::Schema object before it is turned into SQL fed to the database. If you want to execute post-deploy statements which can not be generated by SQL::Translator, the currently suggested method is to overload "deploy" and use dbh_do.



Arguments: @connectinfo
Return Value: $new_schema

Creates and returns a new Schema object. The connection info set on it is used to create a new instance of the storage backend and set it on the Schema object.

See "connect_info" in DBIx::Class::Storage::DBI for DBI-specific syntax on the @connectinfo argument, or DBIx::Class::Storage in general.

Note that connect_info expects an arrayref of arguments, but connect does not. connect wraps its arguments in an arrayref before passing them to connect_info.


connect is a convenience method. It is equivalent to calling $schema->clone->connection(@connectinfo). To write your own overloaded version, overload "connection" instead.


Arguments: $source_name
Return Value: $resultset
  my $rs = $schema->resultset('DVD');

Returns the DBIx::Class::ResultSet object for the registered source name.


Return Value: @source_names
  my @source_names = $schema->sources;

Lists names of all the sources registered on this Schema object.


Arguments: $source_name
Return Value: $result_source
  my $source = $schema->source('Book');

Returns the DBIx::Class::ResultSource object for the registered source name.


Arguments: $source_name
Return Value: $classname
  my $class = $schema->class('CD');

Retrieves the Result class name for the given source name.


Arguments: $coderef, @coderef_args?
Return Value: The return value of $coderef

Executes $coderef with (optional) arguments @coderef_args atomically, returning its result (if any). Equivalent to calling $schema->storage->txn_do. See "txn_do" in DBIx::Class::Storage for more information.

This interface is preferred over using the individual methods "txn_begin", "txn_commit", and "txn_rollback" below.

WARNING: If you are connected with AutoCommit = 0> the transaction is considered nested, and you will still need to call "txn_commit" to write your changes when appropriate. You will also want to connect with auto_savepoint = 1> to get partial rollback to work, if the storage driver for your database supports it.

Connecting with AutoCommit = 1> is recommended.


Runs txn_scope_guard on the schema's storage. See "txn_scope_guard" in DBIx::Class::Storage.


Begins a transaction (does nothing if AutoCommit is off). Equivalent to calling $schema->storage->txn_begin. See "txn_begin" in DBIx::Class::Storage::DBI for more information.


Commits the current transaction. Equivalent to calling $schema->storage->txn_commit. See "txn_commit" in DBIx::Class::Storage::DBI for more information.


Rolls back the current transaction. Equivalent to calling $schema->storage->txn_rollback. See "txn_rollback" in DBIx::Class::Storage::DBI for more information.


  my $storage = $schema->storage;

Returns the DBIx::Class::Storage object for this Schema. Grab this if you want to turn on SQL statement debugging at runtime, or set the quote character. For the default storage, the documentation can be found in DBIx::Class::Storage::DBI.


Arguments: $source_name, \@data;
Return value: \@$objects | nothing

Pass this method a resultsource name, and an arrayref of arrayrefs. The arrayrefs should contain a list of column names, followed by one or many sets of matching data for the given columns.

In void context, insert_bulk in DBIx::Class::Storage::DBI is used to insert the data, as this is a fast method. However, insert_bulk currently assumes that your datasets all contain the same type of values, using scalar references in a column in one row, and not in another will probably not work.

Otherwise, each set of data is inserted into the database using "create" in DBIx::Class::ResultSet, and a arrayref of the resulting row objects is returned.


  $schema->populate('Artist', [
    [ qw/artistid name/ ],
    [ 1, 'Popular Band' ],
    [ 2, 'Indie Band' ],

Since wantarray context is basically the same as looping over $rs->create(...) you won't see any performance benefits and in this case the method is more for convenience. Void context sends the column information directly to storage using <DBI>s bulk insert method. So the performance will be much better for storages that support this method.

Because of this difference in the way void context inserts rows into your database you need to note how this will effect any loaded components that override or augment insert. For example if you are using a component such as DBIx::Class::UUIDColumns to populate your primary keys you MUST use wantarray context if you want the PKs automatically created.


Arguments: @args
Return Value: $new_schema

Similar to "connect" except sets the storage object and connection data in-place on the Schema class. You should probably be calling "connect" to get a proper Schema object instead.


Overload connection to change the behaviour of connect.


Arguments: $target_namespace, $additional_base_class?
Retur Value: $new_schema

For each DBIx::Class::ResultSource in the schema, this method creates a class in the target namespace (e.g. $target_namespace::CD, $target_namespace::Artist) that inherits from the corresponding classes attached to the current schema.

It also attaches a corresponding DBIx::Class::ResultSource object to the new $schema object. If $additional_base_class is given, the new composed classes will inherit from first the corresponding classe from the current schema then the base class.

For example, for a schema with My::Schema::CD and My::Schema::Artist classes,

  $schema->compose_namespace('My::DB', 'Base::Class');
  print join (', ', @My::DB::CD::ISA) . "\n";
  print join (', ', @My::DB::Artist::ISA) ."\n";

will produce the output

  My::Schema::CD, Base::Class
  My::Schema::Artist, Base::Class


Creates a new savepoint (does nothing outside a transaction). Equivalent to calling $schema->storage->svp_begin. See "svp_begin" in DBIx::Class::Storage::DBI for more information.


Releases a savepoint (does nothing outside a transaction). Equivalent to calling $schema->storage->svp_release. See "svp_release" in DBIx::Class::Storage::DBI for more information.


Rollback to a savepoint (does nothing outside a transaction). Equivalent to calling $schema->storage->svp_rollback. See "svp_rollback" in DBIx::Class::Storage::DBI for more information.


Return Value: $new_schema

Clones the schema and its associated result_source objects and returns the copy.


Arguments: $message

Throws an exception. Defaults to using Carp::Clan to report errors from user's perspective. See "exception_action" for details on overriding this method's behavior. If "stacktrace" is turned on, throw_exception's default behavior will provide a detailed stack trace.


Arguments: \%sqlt_args, $dir

Attempts to deploy the schema to the current storage using SQL::Translator.

See "METHODS" in SQL::Translator for a list of values for \%sqlt_args. The most common value for this would be { add_drop_table => 1 } to have the SQL produced include a DROP TABLE statement for each table created. For quoting purposes supply quote_table_names and quote_field_names.

Additionally, the DBIx::Class parser accepts a sources parameter as a hash ref or an array ref, containing a list of source to deploy. If present, then only the sources listed will get deployed. Furthermore, you can use the add_fk_index parser parameter to prevent the parser from creating an index for each FK.


Arguments: See "deployment_statements" in DBIx::Class::Storage::DBI
Return value: $listofstatements

A convenient shortcut to $self->storage->deployment_statements($self, @args). Returns the SQL statements used by "deploy" and "deploy" in DBIx::Class::Schema::Storage.

create_ddl_dir (EXPERIMENTAL)

Arguments: See "create_ddl_dir" in DBIx::Class::Storage::DBI

A convenient shortcut to $self->storage->create_ddl_dir($self, @args).

Creates an SQL file based on the Schema, for each of the specified database types, in the given directory.


Arguments: $database-type, $version, $directory, $preversion
Return value: $normalised_filename
  my $filename = $table->ddl_filename($type, $version, $dir, $preversion)

This method is called by create_ddl_dir to compose a file name out of the supplied directory, database type and version number. The default file name format is: $dir$schema-$version-$type.sql.

You may override this method in your schema if you wish to use a different format.


 Prior to DBIx::Class version 0.08100 this method had a different signature:

    my $filename = $table->ddl_filename($type, $dir, $version, $preversion)

 In recent versions variables $dir and $version were reversed in order to
 bring the signature in line with other Schema/Storage methods. If you 
 really need to maintain backward compatibility, you can do the following
 in any overriding methods:

    ($dir, $version) = ($version, $dir) if ($DBIx::Class::VERSION < 0.08100);


Provided as the recommended way of thawing schema objects. You can call Storable::thaw directly if you wish, but the thawed objects will not have a reference to any schema, so are rather useless


This doesn't actualy do anything more than call "freeze" in Storable, it is just provided here for symetry.


Recommeneded way of dcloning objects. This is needed to properly maintain references to the schema object (which itself is not cloned.)


Returns the current schema class' $VERSION in a normalised way.


Arguments: $moniker, $component_class

This method is called by "load_namespaces" and "load_classes" to install the found classes into your Schema. You should be using those instead of this one.

You will only need this method if you have your Result classes in files which are not named after the packages (or all in the same file). You may also need it to register classes at runtime.

Registers a class which isa DBIx::Class::ResultSourceProxy. Equivalent to calling:

  $schema->register_source($moniker, $component_class->result_source_instance);


Arguments: $moniker, $result_source

This method is called by "register_class".

Registers the DBIx::Class::ResultSource in the schema with the given moniker.


Arguments: $moniker, $result_source

As "register_source" but should be used if the result class already has a source and you want to register an extra one.

compose_connection (DEPRECATED)

Arguments: $target_namespace, @db_info
Return Value: $new_schema

DEPRECATED. You probably wanted compose_namespace.

Actually, you probably just wanted to call connect.


Matt S. Trout <>


You may distribute this code under the same terms as Perl itself.