[Dbix-class] can we store the DBIx::Class into cache system?
Hi, all, I have a question about the how to save the DBIx::Class object into cache system, example: we use CHI, my question is: just save DBIx::Class into Cache system like other data type? I mean just like array, hash, scalar? don't need do other things? so, the DBIx::Class will be lost database connection and lost data? thanks Mike.G ___ 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/dbix-class@lists.scsys.co.uk
RE: [Dbix-class] Unknown column in 'group statement'
Hi Peter, -Original Message- From: Peter Rabbitson [mailto:rabbit+d...@rabbit.us] Sent: 23 July 2009 22:45 To: DBIx::Class user and developer list Subject: Re: [Dbix-class] Unknown column in 'group statement' On Thu, Jul 23, 2009 at 04:19:01PM +0100, Anthony Gladdish wrote: Hi, Using: Perl 5.10. DBIC 0.08108. 1. My result source tables with relationships are: Event.pm: __PACKAGE__-set_primary_key('id'); __PACKAGE__-belongs_to( 'leader' = 'Trainer' ) ; __PACKAGE__-has_many( 'assistingtrainers' = 'AssistingTrainer', 'event' ); __PACKAGE__-many_to_many('assistants' = 'assistingtrainers', 'trainers' ); AssistingTrainer.pm: __PACKAGE__-set_primary_key('event','trainer'); __PACKAGE__-belongs_to( 'events' = 'Event', 'event' ); __PACKAGE__-belongs_to( 'trainers' = 'Trainer', 'trainer' ); Trainer.pm: __PACKAGE__-set_primary_key('id'); 2. My ResultSet::Event.pm methods: sub goEvents { my $self = shift; my $where = {}; $where-{status} = { '' = 2 }; return $self-search($where); } sub events_during_lookback_period { my $self = shift; my $default_lookback_period = DateTime::Duration-new( years = 3 ); my $lookback_period = shift || $default_lookback_period; my $earliest_date = DateTime-now() - $lookback_period; return $self-goEvents()-search_rs( { -and = [ start = { '', DateTime-now() }, start = { '', $earliest_date }, ], } ); } sub distinct_assistants { my $self = shift; my $assistants = []; my $rs = $self-search_rs( undef, { columns = [ qw/me.id/ ], distinct = 1, prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, ); while ( my $e = $rs-next() ) { foreach my $at ( $e-assistingtrainers() ) { push( @$assistants, $at-trainers ); } } return $assistants; } This is a design lapse - we married distinct with group_by too early in the code and what you are seeing is the fallout (i.e. distinct applies to all columns not just the ones you are trying to select). I will have to discuss this with the rest of the developers before we solve it, but for the time being simply change distinct = 1 to group_by = [ qw/me.id/ ] The prefetch will keep working as expected. Cheers This works, but only if I omit columns = too, i.e.: { group_by = [ qw/me.id/ ], prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, I'm assuming this is ok, although I would have thought columns and group_by would be needed together (correct me if I'm wrong), and doing this: { columns = [ qw/me.id/ ], group_by = [ qw/me.id/ ], prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, ... produces error: DBIx::Class::ResultSet::next(): DBI Exception: DBD::mysql::st execute failed: Unknown column 'start' in 'where clause' [for Statement SELECT me.id, assistingtrainers.event, assistingtrainers.trainer, trainers.id, trainers.name, trainers.email, trainers.phone, trainers.initials, trainers.loginId, trainers.password FROM (SELECT me.id FROM event me WHERE ( ( ( start ? AND start ? ) AND status ? ) ) GROUP BY me.id) me LEFT JOIN trainer_assist assistingtrainers ON assistingtrainers.event = me.id LEFT JOIN trainer trainers ON trainers.id = assistingtrainers.trainer WHERE ( ( ( start ? AND start ? ) AND status ? ) ) ORDER BY assistingtrainers.event with ParamValues: 0='2009-07-24T09:46:14', 1='2006-07-24T09:46:14', 2=2, 3='2009-07-24T09:46:14', 4='2006-07-24T09:46:14', 5=2] Thanks for your help, Anthony ___ 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/dbix-class@lists.scsys.co.uk
Re: [Dbix-class] Relations for multi-class object inflation from one table
On Thu, Jul 23, 2009 at 4:12 PM, Oleg Kostyukcub.ua...@gmail.com wrote: Hello all, I read about Dynamic Sub-classing DBIx::Class proxy classes in DBIx::Class::Manual::Cookbook, and have question. package My::Schema::Result::Company; use base qw/DBIx::Class/; __PACKAGE__-load_components(qw/Core/); __PACKAGE__-table('companies'); __PACKAGE__-add_columns(qw/company_id ../); __PACKAGE__-set_primary_key('company_id'); __PACKAGE__-has_many('admins', 'My::Schema::Result::User::Admin', ); __PACKAGE__-has_many('users', 'My::Schema::Result::User', ); __PACKAGE__-has_many('all_users', 'My::Schema::Result::User', ); My questions is: 1) what should be instead of '' in has_many() calls above? I think, this should be something like this (for first has_many): { 'foreign.company_id' = 'self.company_id', 'foreign.admin' = 1 } - is this correct? You can't define rels using constants in FK definitions like 'foreign.admin' = 1. You have to define just the 'all_users' rel with { 'foreign.company_id' = 'self.company_id' } and create result class methods for 'users' and 'admins' like: sub users_rs { my $self = shift; return $self-all_users_rs({ admin = '0' }); } sub users { my $self = shift; my $rs = $self-users_rs; return wantarray ? $rs-all : $rs; } sub admins_rs { my $self = shift; return $self-all_users_rs({ admin = '1' }); } sub admins { my $self = shift; my $rs = $self-admins_rs; return wantarray ? $rs-all : $rs; } Or you can create a base result class called Person, then Admin and User classes as a views (CREATE VIEW) over Person, thus you can define the rels like: __PACKAGE__-has_many('admins', 'My::Schema::Result::User::Admin', { 'foreign.company_id' = 'self.company_id' }); __PACKAGE__-has_many('users', 'My::Schema::Result::User', { 'foreign.company_id' = 'self.company_id' }); __PACKAGE__-has_many('all_users', 'My::Schema::Result::Person', { 'foreign.company_id' = 'self.company_id' }); 3) do I need write belongs_to(company) in User::Admin too, or only in User will be enough? Just in User should be enough. -- wallace reis/wreis Catalyst and DBIx::Class consultancy with a clue Software Developer and a commit bit: http://shadowcat.co.uk/catalyst/ Shadowcat Systems Limited http://www.shadowcat.co.uk http://www.linkedin.com/in/wallacereis ___ 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/dbix-class@lists.scsys.co.uk
Re: [Dbix-class] Unknown column in 'group statement'
On Fri, Jul 24, 2009 at 11:00:50AM +0100, Anthony Gladdish wrote: Hi Peter, -Original Message- From: Peter Rabbitson [mailto:rabbit+d...@rabbit.us] Sent: 23 July 2009 22:45 To: DBIx::Class user and developer list Subject: Re: [Dbix-class] Unknown column in 'group statement' On Thu, Jul 23, 2009 at 04:19:01PM +0100, Anthony Gladdish wrote: Hi, Using: Perl 5.10. DBIC 0.08108. 1. My result source tables with relationships are: Event.pm: __PACKAGE__-set_primary_key('id'); __PACKAGE__-belongs_to( 'leader' = 'Trainer' ) ; __PACKAGE__-has_many( 'assistingtrainers' = 'AssistingTrainer', 'event' ); __PACKAGE__-many_to_many('assistants' = 'assistingtrainers', 'trainers' ); AssistingTrainer.pm: __PACKAGE__-set_primary_key('event','trainer'); __PACKAGE__-belongs_to( 'events' = 'Event', 'event' ); __PACKAGE__-belongs_to( 'trainers' = 'Trainer', 'trainer' ); Trainer.pm: __PACKAGE__-set_primary_key('id'); 2. My ResultSet::Event.pm methods: sub goEvents { my $self = shift; my $where = {}; $where-{status} = { '' = 2 }; return $self-search($where); } sub events_during_lookback_period { my $self = shift; my $default_lookback_period = DateTime::Duration-new( years = 3 ); my $lookback_period = shift || $default_lookback_period; my $earliest_date = DateTime-now() - $lookback_period; return $self-goEvents()-search_rs( { -and = [ start = { '', DateTime-now() }, start = { '', $earliest_date }, ], } ); } sub distinct_assistants { my $self = shift; my $assistants = []; my $rs = $self-search_rs( undef, { columns = [ qw/me.id/ ], distinct = 1, prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, ); while ( my $e = $rs-next() ) { foreach my $at ( $e-assistingtrainers() ) { push( @$assistants, $at-trainers ); } } return $assistants; } This is a design lapse - we married distinct with group_by too early in the code and what you are seeing is the fallout (i.e. distinct applies to all columns not just the ones you are trying to select). I will have to discuss this with the rest of the developers before we solve it, but for the time being simply change distinct = 1 to group_by = [ qw/me.id/ ] The prefetch will keep working as expected. Cheers This works, but only if I omit columns = too, i.e.: { group_by = [ qw/me.id/ ], prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, I'm assuming this is ok, although I would have thought columns and group_by would be needed together (correct me if I'm wrong), and doing this: { columns = [ qw/me.id/ ], group_by = [ qw/me.id/ ], prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, ... produces error: DBIx::Class::ResultSet::next(): DBI Exception: DBD::mysql::st execute failed: Unknown column 'start' in 'where clause' [for Statement SELECT me.id, assistingtrainers.event, assistingtrainers.trainer, trainers.id, trainers.name, trainers.email, trainers.phone, trainers.initials, trainers.loginId, trainers.password FROM (SELECT me.id FROM event me WHERE ( ( ( start ? AND start ? ) AND status ? ) ) GROUP BY me.id) me LEFT JOIN trainer_assist assistingtrainers ON assistingtrainers.event = me.id LEFT JOIN trainer trainers ON trainers.id = assistingtrainers.trainer WHERE ( ( ( start ? AND start ? ) AND status ? ) ) ORDER BY assistingtrainers.event with ParamValues: 0='2009-07-24T09:46:14', 1='2006-07-24T09:46:14', 2=2, 3='2009-07-24T09:46:14', 4='2006-07-24T09:46:14', 5=2] This unfortunately is a limitation which can not be worked around easily. The problem is simple - there is no usable introspection of the WHERE condition, thus there is no sane way to tell *what* the WHERE limits on. In your case WHERE is necessary only on the inner query, as it works on me.* columns only. However since I have no reliable way of determining this, I err on the safe side, and include the WHERE in the inner and outer query[1]. The right way to solve this is to bug ash to finish SQLA2 :) However in the meantime you can trick this with distinct as such: { select = [ { distinct = 'me.id' } ], as = [qw/id/], prefetch = [ { 'assistingtrainers' = 'trainers' } ], join = [ { 'assistingtrainers' = 'trainers' } ], }, HTH [1] The reason the double WHERE is