On Jan 14, 2008, at 6:40 AM, Carl Franks wrote:
I'm unable to use create() to create a new record with a has_one
relationship.
The docs (ResultSet/create) say: for has_one rels, "pass an arrayref
of hashrefs".
This looks like a doc bug, I don't see any reason to be passing an
arrayref when has_one can only refer to one object. I suspect it was
supposed to instruct you to use a hashref for belongs_to/might_have/
has_one (relationships where accessor = 'single') and an arrayref for
has_many/many_to_many (relationships where accessor = 'multi').
If I pass an arrayref containing a hashref, like so:
$rs->create({
text_col => 'filler2',
user => [ { name => 'foo' } ]
} );
I get this error message - I don't know where the "NAME" in the SQL
comes from:
SELECT me.id, me.master, me.name FROM user me WHERE ( ( ( ( me.id NAME
? ) ) ) ): 'foo'
DBI Exception: DBD::SQLite::db prepare_cached failed: near "NAME":
syntax error(1) at dbdimp.c line 271 [for Statement "SELECT me.id,
me.master, me.name FROM user me WHERE ( ( ( ( me.id NAME ? ) ) ) )"]
at /opt/perl-5.8.8/lib/site_perl/5.8.8/DBIx/Class/Schema.pm line 945
Looks like it's the result of ->find interpreting your user wrong
(because it didn't actually expect to see an arrayref there), and
trying to turn that into part of the query.
If I get rid of the arrayref, and just pass a hashref for the rel
data, like so:
$rs->create({
text_col => 'filler2',
user => { name => 'foo' }
} );
The inserts work, but the related (user) table doesn't get the
belongs_to id set.
Setting DBIC_TRACE shows the SQL being generated:
SELECT me.id, me.master, me.name FROM user me WHERE ( me.name = ? ):
'foo'
BEGIN WORK
INSERT INTO master (id, text_col) VALUES (?, ?): 'NULL', 'filler2'
SELECT COUNT( * ) FROM user me WHERE ( name = ? ): 'foo'
INSERT INTO user (name) VALUES (?): 'foo'
COMMIT
I don't know why it's trying to retrieve a has_one rel (user) before
the main entry (master) is even inserted.
Internally, what it's doing when you pass a hashref for a 'single'
object (belongs_to/might_have/has_one) is something like this:
my $master = $rs->new( { text_col => 'filler2' } );
my $user = $master->find_or_new_related( { name => 'foo' } );
$master->user( $user );
$master->insert;
So the initial select is from the find_or_new_related attempting to
figure out if it already exists or not.
The "SELECT COUNT(*)" also looks suspicious - is that intended?
This is the result of ->insert() using ->count() to determine if the
related object that it's about to insert into the database is already
in storage or not. I think what's happening is that one of your
relationships is causing the initial select, and the other one is
causing the count.
Am I using this wrong?
As far as I can tell it _should_ work, but I generally opt to create
the related objects directly, which may help to avoid the problem you
are encountering...
$rs->create({
text_col => 'filler2',
user => $schema->resultset( 'User' )->find_or_create( name =>
'foo' ),
} );
I'm using DBIx::Class 0.08008, DBD::SQLite 1.14, sqlite 3.4.2-3.
Copied below is the 2 relevant table classes (with irrelevant cols and
rels deleted for clarity).
package MySchema::Master;
use base 'DBIx::Class';
__PACKAGE__->load_components(qw/
InflateColumn::DateTime Core
/);
__PACKAGE__->table("master");
__PACKAGE__->add_columns(
id => { data_type => "INTEGER" },
text_col => { data_type => "TEXT" },
);
__PACKAGE__->set_primary_key("id");
__PACKAGE__->has_one( user => 'MySchema::User', 'master' );
1;
package MySchema::User;
use base 'DBIx::Class';
__PACKAGE__->load_components(qw/ Core /);
__PACKAGE__->table("user");
__PACKAGE__->add_columns(
id => { data_type => "INTEGER" },
master => { data_type => "INTEGER" },
name => { data_type => "TEXT" },
);
__PACKAGE__->set_primary_key("id");
__PACKAGE__->belongs_to( master => 'MySchema::Master', 'id' );
1;
_______________________________________________
List: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/dbix-class
IRC: irc.perl.org#dbix-class
SVN: http://dev.catalyst.perl.org/repos/bast/DBIx-Class/
Searchable Archive: http://www.grokbase.com/group/[EMAIL PROTECTED]
--
Jason Kohles, RHCA RHCDS RHCE
[EMAIL PROTECTED] - http://www.jasonkohles.com/
"A witty saying proves nothing." -- Voltaire
_______________________________________________
List: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/dbix-class
IRC: irc.perl.org#dbix-class
SVN: http://dev.catalyst.perl.org/repos/bast/DBIx-Class/
Searchable Archive: http://www.grokbase.com/group/[EMAIL PROTECTED]
