1 package DBIx::Class::Storage::DBI;
2 # -*- mode: cperl; cperl-indent-level: 2 -*-
4 use base 'DBIx::Class::Storage';
8 use Carp::Clan qw/^DBIx::Class/;
10 use DBIx::Class::Storage::DBI::Cursor;
11 use DBIx::Class::Storage::Statistics;
15 __PACKAGE__->mk_group_accessors('simple' =>
16 qw/_connect_info _dbi_connect_info _dbh _sql_maker _sql_maker_opts
17 _conn_pid _conn_tid transaction_depth _dbh_autocommit _on_connect_do
18 _on_disconnect_do _on_connect_do_store _on_disconnect_do_store
22 # the values for these accessors are picked out (and deleted) from
23 # the attribute hashref passed to connect_info
24 my @storage_options = qw/
25 on_connect_call on_disconnect_call disable_sth_caching unsafe auto_savepoint
27 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
30 # default cursor class, overridable in connect_info attributes
31 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
33 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class/);
34 __PACKAGE__->sql_maker_class('DBIx::Class::SQLAHacks');
39 DBIx::Class::Storage::DBI - DBI storage handler
43 my $schema = MySchema->connect('dbi:SQLite:my.db');
45 $schema->storage->debug(1);
46 $schema->dbh_do("DROP TABLE authors");
48 $schema->resultset('Book')->search({
49 written_on => $schema->storage->datetime_parser(DateTime->now)
54 This class represents the connection to an RDBMS via L<DBI>. See
55 L<DBIx::Class::Storage> for general information. This pod only
56 documents DBI-specific methods and behaviors.
63 my $new = shift->next::method(@_);
65 $new->transaction_depth(0);
66 $new->_sql_maker_opts({});
67 $new->{savepoints} = [];
68 $new->{_in_dbh_do} = 0;
76 This method is normally called by L<DBIx::Class::Schema/connection>, which
77 encapsulates its argument list in an arrayref before passing them here.
79 The argument list may contain:
85 The same 4-element argument set one would normally pass to
86 L<DBI/connect>, optionally followed by
87 L<extra attributes|/DBIx::Class specific connection attributes>
88 recognized by DBIx::Class:
90 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
94 A single code reference which returns a connected
95 L<DBI database handle|DBI/connect> optionally followed by
96 L<extra attributes|/DBIx::Class specific connection attributes> recognized
99 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
103 A single hashref with all the attributes and the dsn/user/password
106 $connect_info_args = [{
114 This is particularly useful for L<Catalyst> based applications, allowing the
115 following config (L<Config::General> style):
120 dsn dbi:mysql:database=test
129 Please note that the L<DBI> docs recommend that you always explicitly
130 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
131 recommends that it be set to I<1>, and that you perform transactions
132 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
133 to I<1> if you do not do explicitly set it to zero. This is the default
134 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
136 =head3 DBIx::Class specific connection attributes
138 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
139 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
140 the following connection options. These options can be mixed in with your other
141 L<DBI> connection attributes, or placed in a seperate hashref
142 (C<\%extra_attributes>) as shown above.
144 Every time C<connect_info> is invoked, any previous settings for
145 these options will be cleared before setting the new ones, regardless of
146 whether any options are specified in the new C<connect_info>.
153 Specifies things to do immediately after connecting or re-connecting to
154 the database. Its value may contain:
160 This contains one SQL statement to execute.
162 =item an array reference
164 This contains SQL statements to execute in order. Each element contains
165 a string or a code reference that returns a string.
167 =item a code reference
169 This contains some code to execute. Unlike code references within an
170 array reference, its return value is ignored.
174 =item on_disconnect_do
176 Takes arguments in the same form as L</on_connect_do> and executes them
177 immediately before disconnecting from the database.
179 Note, this only runs if you explicitly call L</disconnect> on the
182 =item on_connect_call
184 A more generalized form of L</on_connect_do> that calls the specified
185 C<connect_call_METHOD> methods in your storage driver.
187 on_connect_do => 'select 1'
191 on_connect_call => [ [ do_sql => 'select 1' ] ]
193 Its values may contain:
199 Will call the C<connect_call_METHOD> method.
201 =item a code reference
203 Will execute C<< $code->($storage) >>
205 =item an array reference
207 Each value can be a method name or code reference.
209 =item an array of arrays
211 For each array, the first item is taken to be the C<connect_call_> method name
212 or code reference, and the rest are parameters to it.
216 Some predefined storage methods you may use:
222 Executes a SQL string or a code reference that returns a SQL string. This is
223 what L</on_connect_do> and L</on_disconnect_do> use.
231 Will execute the scalar as SQL.
235 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
236 attributes hashref and bind values.
238 =item a code reference
240 Will execute C<< $code->($storage) >> and execute the return array refs as
247 Execute any statements necessary to initialize the database session to return
248 and accept datetime/timestamp values used with
249 L<DBIx::Class::InflateColumn::DateTime>.
251 Only necessary for some databases, see your specific storage driver for
252 implementation details.
256 =item on_disconnect_call
258 Takes arguments in the same form as L</on_connect_call> and executes them
259 immediately before disconnecting from the database.
261 Calls the C<disconnect_call_METHOD> methods as opposed to the
262 C<connect_call_METHOD> methods called by L</on_connect_call>.
264 Note, this only runs if you explicitly call L</disconnect> on the
267 =item disable_sth_caching
269 If set to a true value, this option will disable the caching of
270 statement handles via L<DBI/prepare_cached>.
274 Sets the limit dialect. This is useful for JDBC-bridge among others
275 where the remote SQL-dialect cannot be determined by the name of the
276 driver alone. See also L<SQL::Abstract::Limit>.
280 Specifies what characters to use to quote table and column names. If
281 you use this you will want to specify L</name_sep> as well.
283 C<quote_char> expects either a single character, in which case is it
284 is placed on either side of the table/column name, or an arrayref of length
285 2 in which case the table/column name is placed between the elements.
287 For example under MySQL you should use C<< quote_char => '`' >>, and for
288 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
292 This only needs to be used in conjunction with C<quote_char>, and is used to
293 specify the charecter that seperates elements (schemas, tables, columns) from
294 each other. In most cases this is simply a C<.>.
296 The consequences of not supplying this value is that L<SQL::Abstract>
297 will assume DBIx::Class' uses of aliases to be complete column
298 names. The output will look like I<"me.name"> when it should actually
303 This Storage driver normally installs its own C<HandleError>, sets
304 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
305 all database handles, including those supplied by a coderef. It does this
306 so that it can have consistent and useful error behavior.
308 If you set this option to a true value, Storage will not do its usual
309 modifications to the database handle's attributes, and instead relies on
310 the settings in your connect_info DBI options (or the values you set in
311 your connection coderef, in the case that you are connecting via coderef).
313 Note that your custom settings can cause Storage to malfunction,
314 especially if you set a C<HandleError> handler that suppresses exceptions
315 and/or disable C<RaiseError>.
319 If this option is true, L<DBIx::Class> will use savepoints when nesting
320 transactions, making it possible to recover from failure in the inner
321 transaction without having to abort all outer transactions.
325 Use this argument to supply a cursor class other than the default
326 L<DBIx::Class::Storage::DBI::Cursor>.
330 Some real-life examples of arguments to L</connect_info> and
331 L<DBIx::Class::Schema/connect>
333 # Simple SQLite connection
334 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
337 ->connect_info([ sub { DBI->connect(...) } ]);
339 # A bit more complicated
346 { quote_char => q{"}, name_sep => q{.} },
350 # Equivalent to the previous example
356 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
360 # Same, but with hashref as argument
361 # See parse_connect_info for explanation
364 dsn => 'dbi:Pg:dbname=foo',
366 password => 'my_pg_password',
373 # Subref + DBIx::Class-specific connection options
376 sub { DBI->connect(...) },
380 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
381 disable_sth_caching => 1,
391 my ($self, $info_arg) = @_;
393 return $self->_connect_info if !$info_arg;
395 my @args = @$info_arg; # take a shallow copy for further mutilation
396 $self->_connect_info([@args]); # copy for _connect_info
399 # combine/pre-parse arguments depending on invocation style
402 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
403 %attrs = %{ $args[1] || {} };
406 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
407 %attrs = %{$args[0]};
409 for (qw/password user dsn/) {
410 unshift @args, delete $attrs{$_};
413 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
415 % { $args[3] || {} },
416 % { $args[4] || {} },
418 @args = @args[0,1,2];
421 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
422 # the new set of options
423 $self->_sql_maker(undef);
424 $self->_sql_maker_opts({});
427 for my $storage_opt (@storage_options, 'cursor_class') { # @storage_options is declared at the top of the module
428 if(my $value = delete $attrs{$storage_opt}) {
429 $self->$storage_opt($value);
432 for my $sql_maker_opt (qw/limit_dialect quote_char name_sep/) {
433 if(my $opt_val = delete $attrs{$sql_maker_opt}) {
434 $self->_sql_maker_opts->{$sql_maker_opt} = $opt_val;
437 for my $connect_do_opt (qw/on_connect_do on_disconnect_do/) {
438 if(my $opt_val = delete $attrs{$connect_do_opt}) {
439 $self->$connect_do_opt($opt_val);
444 %attrs = () if (ref $args[0] eq 'CODE'); # _connect() never looks past $args[0] in this case
446 $self->_dbi_connect_info([@args, keys %attrs ? \%attrs : ()]);
447 $self->_connect_info;
452 This method is deprecated in favour of setting via L</connect_info>.
458 $self->_setup_connect_do(on_connect_do => @_);
461 =head2 on_disconnect_do
463 This method is deprecated in favour of setting via L</connect_info>.
467 sub on_disconnect_do {
469 $self->_setup_connect_do(on_disconnect_do => @_);
472 sub _setup_connect_do {
473 my ($self, $opt) = (shift, shift);
475 my $accessor = "_$opt";
476 my $store = "_${opt}_store";
478 return $self->$accessor if not @_;
482 if (not defined $val) {
483 $self->$accessor(undef);
484 $self->$store(undef);
491 push @store, [ 'do_sql', $val ];
492 } elsif (ref($val) eq 'CODE') {
494 } elsif (ref($val) eq 'ARRAY') {
495 push @store, map [ 'do_sql', $_ ], @$val;
497 $self->throw_exception("Invalid type for $opt ".ref($val));
500 $self->$store(\@store);
501 $self->$accessor($val);
506 Arguments: ($subref | $method_name), @extra_coderef_args?
508 Execute the given $subref or $method_name using the new exception-based
509 connection management.
511 The first two arguments will be the storage object that C<dbh_do> was called
512 on and a database handle to use. Any additional arguments will be passed
513 verbatim to the called subref as arguments 2 and onwards.
515 Using this (instead of $self->_dbh or $self->dbh) ensures correct
516 exception handling and reconnection (or failover in future subclasses).
518 Your subref should have no side-effects outside of the database, as
519 there is the potential for your subref to be partially double-executed
520 if the database connection was stale/dysfunctional.
524 my @stuff = $schema->storage->dbh_do(
526 my ($storage, $dbh, @cols) = @_;
527 my $cols = join(q{, }, @cols);
528 $dbh->selectrow_array("SELECT $cols FROM foo");
539 my $dbh = $self->_dbh;
541 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
542 || $self->{transaction_depth};
544 local $self->{_in_dbh_do} = 1;
547 my $want_array = wantarray;
550 $self->_verify_pid if $dbh;
552 $self->_populate_dbh;
557 @result = $self->$code($dbh, @_);
559 elsif(defined $want_array) {
560 $result[0] = $self->$code($dbh, @_);
563 $self->$code($dbh, @_);
568 if(!$exception) { return $want_array ? @result : $result[0] }
570 $self->throw_exception($exception) if $self->connected;
572 # We were not connected - reconnect and retry, but let any
573 # exception fall right through this time
574 $self->_populate_dbh;
575 $self->$code($self->_dbh, @_);
578 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
579 # It also informs dbh_do to bypass itself while under the direction of txn_do,
580 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
585 ref $coderef eq 'CODE' or $self->throw_exception
586 ('$coderef must be a CODE reference');
588 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
590 local $self->{_in_dbh_do} = 1;
593 my $want_array = wantarray;
598 $self->_verify_pid if $self->_dbh;
599 $self->_populate_dbh if !$self->_dbh;
603 @result = $coderef->(@_);
605 elsif(defined $want_array) {
606 $result[0] = $coderef->(@_);
615 if(!$exception) { return $want_array ? @result : $result[0] }
617 if($tried++ > 0 || $self->connected) {
618 eval { $self->txn_rollback };
619 my $rollback_exception = $@;
620 if($rollback_exception) {
621 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
622 $self->throw_exception($exception) # propagate nested rollback
623 if $rollback_exception =~ /$exception_class/;
625 $self->throw_exception(
626 "Transaction aborted: ${exception}. "
627 . "Rollback failed: ${rollback_exception}"
630 $self->throw_exception($exception)
633 # We were not connected, and was first try - reconnect and retry
635 $self->_populate_dbh;
641 Our C<disconnect> method also performs a rollback first if the
642 database is not in C<AutoCommit> mode.
649 if( $self->connected ) {
650 if (my $connection_call = $self->on_disconnect_call) {
651 $self->_do_connection_actions(disconnect_call_ => $connection_call)
653 if (my $connection_do = $self->_on_disconnect_do_store) {
654 $self->_do_connection_actions(disconnect_call_ => $connection_do)
657 $self->_dbh->rollback unless $self->_dbh_autocommit;
658 $self->_dbh->disconnect;
664 =head2 with_deferred_fk_checks
668 =item Arguments: C<$coderef>
670 =item Return Value: The return value of $coderef
674 Storage specific method to run the code ref with FK checks deferred or
675 in MySQL's case disabled entirely.
679 # Storage subclasses should override this
680 sub with_deferred_fk_checks {
681 my ($self, $sub) = @_;
689 if(my $dbh = $self->_dbh) {
690 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
697 return 0 if !$self->_dbh;
699 return ($dbh->FETCH('Active') && $dbh->ping);
705 # handle pid changes correctly
706 # NOTE: assumes $self->_dbh is a valid $dbh
710 return if defined $self->_conn_pid && $self->_conn_pid == $$;
712 $self->_dbh->{InactiveDestroy} = 1;
719 sub ensure_connected {
722 unless ($self->connected) {
723 $self->_populate_dbh;
729 Returns the dbh - a data base handle of class L<DBI>.
736 $self->ensure_connected;
740 sub _sql_maker_args {
743 return ( bindtype=>'columns', array_datatypes => 1, limit_dialect => $self->dbh, %{$self->_sql_maker_opts} );
748 unless ($self->_sql_maker) {
749 my $sql_maker_class = $self->sql_maker_class;
750 $self->ensure_class_loaded ($sql_maker_class);
751 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
753 return $self->_sql_maker;
760 my @info = @{$self->_dbi_connect_info || []};
761 $self->_dbh($self->_connect(@info));
763 $self->_conn_pid($$);
764 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
766 $self->_determine_driver;
768 # Always set the transaction depth on connect, since
769 # there is no transaction in progress by definition
770 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
772 if (my $connection_call = $self->on_connect_call) {
773 $self->_do_connection_actions(connect_call_ => $connection_call)
775 if (my $connection_do = $self->_on_connect_do_store) {
776 $self->_do_connection_actions(connect_call_ => $connection_do)
780 sub _determine_driver {
783 if (ref $self eq 'DBIx::Class::Storage::DBI') {
786 if ($self->_dbh) { # we are connected
787 $driver = $self->_dbh->{Driver}{Name};
789 # try to use dsn to not require being connected, the driver may still
790 # force a connection in _rebless to determine version
791 ($driver) = $self->_dbi_connect_info->[0] =~ /dbi:([^:]+):/i;
794 if ($self->load_optional_class("DBIx::Class::Storage::DBI::${driver}")) {
795 bless $self, "DBIx::Class::Storage::DBI::${driver}";
801 sub _do_connection_actions {
803 my $method_prefix = shift;
806 if (not ref($call)) {
807 my $method = $method_prefix . $call;
809 } elsif (ref($call) eq 'CODE') {
811 } elsif (ref($call) eq 'ARRAY') {
812 if (ref($call->[0]) ne 'ARRAY') {
813 $self->_do_connection_actions($method_prefix, $_) for @$call;
815 $self->_do_connection_actions($method_prefix, @$_) for @$call;
818 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
824 sub connect_call_do_sql {
826 $self->_do_query(@_);
829 sub disconnect_call_do_sql {
831 $self->_do_query(@_);
834 # override in db-specific backend when necessary
835 sub connect_call_datetime_setup { 1 }
838 my ($self, $action) = @_;
840 if (ref $action eq 'CODE') {
841 $action = $action->($self);
842 $self->_do_query($_) foreach @$action;
845 # Most debuggers expect ($sql, @bind), so we need to exclude
846 # the attribute hash which is the second argument to $dbh->do
847 # furthermore the bind values are usually to be presented
848 # as named arrayref pairs, so wrap those here too
849 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
850 my $sql = shift @do_args;
851 my $attrs = shift @do_args;
852 my @bind = map { [ undef, $_ ] } @do_args;
854 $self->_query_start($sql, @bind);
855 $self->_dbh->do($sql, $attrs, @do_args);
856 $self->_query_end($sql, @bind);
863 my ($self, @info) = @_;
865 $self->throw_exception("You failed to provide any connection info")
868 my ($old_connect_via, $dbh);
870 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
871 $old_connect_via = $DBI::connect_via;
872 $DBI::connect_via = 'connect';
876 if(ref $info[0] eq 'CODE') {
880 $dbh = DBI->connect(@info);
883 if($dbh && !$self->unsafe) {
884 my $weak_self = $self;
885 Scalar::Util::weaken($weak_self);
886 $dbh->{HandleError} = sub {
888 $weak_self->throw_exception("DBI Exception: $_[0]");
891 croak ("DBI Exception: $_[0]");
894 $dbh->{ShowErrorStatement} = 1;
895 $dbh->{RaiseError} = 1;
896 $dbh->{PrintError} = 0;
900 $DBI::connect_via = $old_connect_via if $old_connect_via;
902 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
905 $self->_dbh_autocommit($dbh->{AutoCommit});
911 my ($self, $name) = @_;
913 $name = $self->_svp_generate_name
914 unless defined $name;
916 $self->throw_exception ("You can't use savepoints outside a transaction")
917 if $self->{transaction_depth} == 0;
919 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
920 unless $self->can('_svp_begin');
922 push @{ $self->{savepoints} }, $name;
924 $self->debugobj->svp_begin($name) if $self->debug;
926 return $self->_svp_begin($name);
930 my ($self, $name) = @_;
932 $self->throw_exception ("You can't use savepoints outside a transaction")
933 if $self->{transaction_depth} == 0;
935 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
936 unless $self->can('_svp_release');
939 $self->throw_exception ("Savepoint '$name' does not exist")
940 unless grep { $_ eq $name } @{ $self->{savepoints} };
942 # Dig through the stack until we find the one we are releasing. This keeps
943 # the stack up to date.
946 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
948 $name = pop @{ $self->{savepoints} };
951 $self->debugobj->svp_release($name) if $self->debug;
953 return $self->_svp_release($name);
957 my ($self, $name) = @_;
959 $self->throw_exception ("You can't use savepoints outside a transaction")
960 if $self->{transaction_depth} == 0;
962 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
963 unless $self->can('_svp_rollback');
966 # If they passed us a name, verify that it exists in the stack
967 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
968 $self->throw_exception("Savepoint '$name' does not exist!");
971 # Dig through the stack until we find the one we are releasing. This keeps
972 # the stack up to date.
973 while(my $s = pop(@{ $self->{savepoints} })) {
974 last if($s eq $name);
976 # Add the savepoint back to the stack, as a rollback doesn't remove the
977 # named savepoint, only everything after it.
978 push(@{ $self->{savepoints} }, $name);
980 # We'll assume they want to rollback to the last savepoint
981 $name = $self->{savepoints}->[-1];
984 $self->debugobj->svp_rollback($name) if $self->debug;
986 return $self->_svp_rollback($name);
989 sub _svp_generate_name {
992 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
997 $self->ensure_connected();
998 if($self->{transaction_depth} == 0) {
999 $self->debugobj->txn_begin()
1001 # this isn't ->_dbh-> because
1002 # we should reconnect on begin_work
1003 # for AutoCommit users
1004 $self->dbh->begin_work;
1005 } elsif ($self->auto_savepoint) {
1008 $self->{transaction_depth}++;
1013 if ($self->{transaction_depth} == 1) {
1014 my $dbh = $self->_dbh;
1015 $self->debugobj->txn_commit()
1018 $self->{transaction_depth} = 0
1019 if $self->_dbh_autocommit;
1021 elsif($self->{transaction_depth} > 1) {
1022 $self->{transaction_depth}--;
1024 if $self->auto_savepoint;
1030 my $dbh = $self->_dbh;
1032 if ($self->{transaction_depth} == 1) {
1033 $self->debugobj->txn_rollback()
1035 $self->{transaction_depth} = 0
1036 if $self->_dbh_autocommit;
1039 elsif($self->{transaction_depth} > 1) {
1040 $self->{transaction_depth}--;
1041 if ($self->auto_savepoint) {
1042 $self->svp_rollback;
1047 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1052 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1053 $error =~ /$exception_class/ and $self->throw_exception($error);
1054 # ensure that a failed rollback resets the transaction depth
1055 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1056 $self->throw_exception($error);
1060 # This used to be the top-half of _execute. It was split out to make it
1061 # easier to override in NoBindVars without duping the rest. It takes up
1062 # all of _execute's args, and emits $sql, @bind.
1063 sub _prep_for_execute {
1064 my ($self, $op, $extra_bind, $ident, $args) = @_;
1066 if( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1067 $ident = $ident->from();
1070 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1073 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1075 return ($sql, \@bind);
1079 sub _fix_bind_params {
1080 my ($self, @bind) = @_;
1082 ### Turn @bind from something like this:
1083 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1085 ### ( "'1'", "'1'", "'3'" )
1088 if ( defined( $_ && $_->[1] ) ) {
1089 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1096 my ( $self, $sql, @bind ) = @_;
1098 if ( $self->debug ) {
1099 @bind = $self->_fix_bind_params(@bind);
1101 $self->debugobj->query_start( $sql, @bind );
1106 my ( $self, $sql, @bind ) = @_;
1108 if ( $self->debug ) {
1109 @bind = $self->_fix_bind_params(@bind);
1110 $self->debugobj->query_end( $sql, @bind );
1115 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1117 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1119 $self->_query_start( $sql, @$bind );
1121 my $sth = $self->sth($sql,$op);
1123 my $placeholder_index = 1;
1125 foreach my $bound (@$bind) {
1126 my $attributes = {};
1127 my($column_name, @data) = @$bound;
1129 if ($bind_attributes) {
1130 $attributes = $bind_attributes->{$column_name}
1131 if defined $bind_attributes->{$column_name};
1134 foreach my $data (@data) {
1135 my $ref = ref $data;
1136 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1138 $sth->bind_param($placeholder_index, $data, $attributes);
1139 $placeholder_index++;
1143 # Can this fail without throwing an exception anyways???
1144 my $rv = $sth->execute();
1145 $self->throw_exception($sth->errstr) if !$rv;
1147 $self->_query_end( $sql, @$bind );
1149 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1154 $self->dbh_do('_dbh_execute', @_)
1158 my ($self, $source, $to_insert) = @_;
1160 my $ident = $source->from;
1161 my $bind_attributes = $self->source_bind_attributes($source);
1163 my $updated_cols = {};
1165 $self->ensure_connected;
1166 foreach my $col ( $source->columns ) {
1167 if ( !defined $to_insert->{$col} ) {
1168 my $col_info = $source->column_info($col);
1170 if ( $col_info->{auto_nextval} ) {
1171 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch( 'nextval', $col_info->{sequence} || $self->_dbh_get_autoinc_seq($self->dbh, $source) );
1176 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
1178 return $updated_cols;
1181 ## Still not quite perfect, and EXPERIMENTAL
1182 ## Currently it is assumed that all values passed will be "normal", i.e. not
1183 ## scalar refs, or at least, all the same type as the first set, the statement is
1184 ## only prepped once.
1186 my ($self, $source, $cols, $data) = @_;
1188 my $table = $source->from;
1189 @colvalues{@$cols} = (0..$#$cols);
1190 my ($sql, @bind) = $self->sql_maker->insert($table, \%colvalues);
1192 $self->_query_start( $sql, @bind );
1193 my $sth = $self->sth($sql);
1195 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1197 ## This must be an arrayref, else nothing works!
1198 my $tuple_status = [];
1200 ## Get the bind_attributes, if any exist
1201 my $bind_attributes = $self->source_bind_attributes($source);
1203 ## Bind the values and execute
1204 my $placeholder_index = 1;
1206 foreach my $bound (@bind) {
1208 my $attributes = {};
1209 my ($column_name, $data_index) = @$bound;
1211 if( $bind_attributes ) {
1212 $attributes = $bind_attributes->{$column_name}
1213 if defined $bind_attributes->{$column_name};
1216 my @data = map { $_->[$data_index] } @$data;
1218 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1219 $placeholder_index++;
1221 my $rv = eval { $sth->execute_array({ArrayTupleStatus => $tuple_status}) };
1224 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1226 $self->throw_exception($sth->errstr || "Unexpected populate error: $err")
1227 if ($i > $#$tuple_status);
1229 require Data::Dumper;
1230 local $Data::Dumper::Terse = 1;
1231 local $Data::Dumper::Indent = 1;
1232 local $Data::Dumper::Useqq = 1;
1233 local $Data::Dumper::Quotekeys = 0;
1235 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1236 $tuple_status->[$i][1],
1237 Data::Dumper::Dumper(
1238 { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) }
1242 $self->throw_exception($sth->errstr) if !$rv;
1244 $self->_query_end( $sql, @bind );
1245 return (wantarray ? ($rv, $sth, @bind) : $rv);
1249 my $self = shift @_;
1250 my $source = shift @_;
1251 my $bind_attributes = $self->source_bind_attributes($source);
1253 return $self->_execute('update' => [], $source, $bind_attributes, @_);
1258 my $self = shift @_;
1259 my $source = shift @_;
1261 my $bind_attrs = $self->source_bind_attributes($source);
1263 return $self->_execute('delete' => [], $source, $bind_attrs, @_);
1266 # We were sent here because the $rs contains a complex search
1267 # which will require a subquery to select the correct rows
1268 # (i.e. joined or limited resultsets)
1270 # Genarating a single PK column subquery is trivial and supported
1271 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1272 # Look at _multipk_update_delete()
1273 sub _subq_update_delete {
1275 my ($rs, $op, $values) = @_;
1277 my $rsrc = $rs->result_source;
1279 # we already check this, but double check naively just in case. Should be removed soon
1280 my $sel = $rs->_resolved_attrs->{select};
1281 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1282 my @pcols = $rsrc->primary_columns;
1283 if (@$sel != @pcols) {
1284 $self->throw_exception (
1285 'Subquery update/delete can not be called on resultsets selecting a'
1286 .' number of columns different than the number of primary keys'
1293 $op eq 'update' ? $values : (),
1294 { $pcols[0] => { -in => $rs->as_query } },
1299 return $self->_multipk_update_delete (@_);
1303 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1304 # resultset update/delete involving subqueries. So by default resort
1305 # to simple (and inefficient) delete_all style per-row opearations,
1306 # while allowing specific storages to override this with a faster
1309 sub _multipk_update_delete {
1310 return shift->_per_row_update_delete (@_);
1313 # This is the default loop used to delete/update rows for multi PK
1314 # resultsets, and used by mysql exclusively (because it can't do anything
1317 # We do not use $row->$op style queries, because resultset update/delete
1318 # is not expected to cascade (this is what delete_all/update_all is for).
1320 # There should be no race conditions as the entire operation is rolled
1323 sub _per_row_update_delete {
1325 my ($rs, $op, $values) = @_;
1327 my $rsrc = $rs->result_source;
1328 my @pcols = $rsrc->primary_columns;
1330 my $guard = $self->txn_scope_guard;
1332 # emulate the return value of $sth->execute for non-selects
1333 my $row_cnt = '0E0';
1335 my $subrs_cur = $rs->cursor;
1336 while (my @pks = $subrs_cur->next) {
1339 for my $i (0.. $#pcols) {
1340 $cond->{$pcols[$i]} = $pks[$i];
1345 $op eq 'update' ? $values : (),
1360 # localization is neccessary as
1361 # 1) there is no infrastructure to pass this around (easy to do, but will wait)
1362 # 2) _select_args sets it and _prep_for_execute consumes it
1363 my $sql_maker = $self->sql_maker;
1364 local $sql_maker->{for};
1366 return $self->_execute($self->_select_args(@_));
1369 sub _select_args_to_query {
1372 # localization is neccessary as
1373 # 1) there is no infrastructure to pass this around (easy to do, but will wait)
1374 # 2) _select_args sets it and _prep_for_execute consumes it
1375 my $sql_maker = $self->sql_maker;
1376 local $sql_maker->{for};
1378 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $order, $rows, $offset)
1379 # = $self->_select_args($ident, $select, $cond, $attrs);
1380 my ($op, $bind, $ident, $bind_attrs, @args) =
1381 $self->_select_args(@_);
1383 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $order, $rows, $offset ]);
1384 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1385 $prepared_bind ||= [];
1388 ? ($sql, $prepared_bind, $bind_attrs)
1389 : \[ "($sql)", @$prepared_bind ]
1394 my ($self, $ident, $select, $where, $attrs) = @_;
1396 my $sql_maker = $self->sql_maker;
1397 my $alias2source = $self->_resolve_ident_sources ($ident);
1399 # calculate bind_attrs before possible $ident mangling
1400 my $bind_attrs = {};
1401 for my $alias (keys %$alias2source) {
1402 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
1403 for my $col (keys %$bindtypes) {
1405 my $fqcn = join ('.', $alias, $col);
1406 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
1408 # so that unqualified searches can be bound too
1409 $bind_attrs->{$col} = $bind_attrs->{$fqcn} if $alias eq 'me';
1414 if ($attrs->{software_limit} ||
1415 $sql_maker->_default_limit_syntax eq "GenericSubQ") {
1416 $attrs->{software_limit} = 1;
1418 $self->throw_exception("rows attribute must be positive if present")
1419 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1421 # MySQL actually recommends this approach. I cringe.
1422 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1424 if ($attrs->{rows} && keys %{$attrs->{collapse}}) {
1425 ($ident, $select, $where, $attrs)
1426 = $self->_adjust_select_args_for_limited_prefetch ($ident, $select, $where, $attrs);
1429 push @limit, $attrs->{rows}, $attrs->{offset};
1434 # This would be the point to deflate anything found in $where
1435 # (and leave $attrs->{bind} intact). Problem is - inflators historically
1436 # expect a row object. And all we have is a resultsource (it is trivial
1437 # to extract deflator coderefs via $alias2source above).
1439 # I don't see a way forward other than changing the way deflators are
1440 # invoked, and that's just bad...
1444 { $attrs->{$_} ? ( $_ => $attrs->{$_} ) : () }
1445 (qw/order_by group_by having _virtual_order_by/ )
1449 $sql_maker->{for} = delete $attrs->{for};
1451 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $order, @limit);
1454 sub _adjust_select_args_for_limited_prefetch {
1455 my ($self, $from, $select, $where, $attrs) = @_;
1457 if ($attrs->{group_by} and @{$attrs->{group_by}}) {
1458 $self->throw_exception ('Prefetch with limit (rows/offset) is not supported on resultsets with a group_by attribute');
1461 $self->throw_exception ('Prefetch with limit (rows/offset) is not supported on resultsets with a custom from attribute')
1462 if (ref $from ne 'ARRAY');
1464 # separate attributes
1465 my $sub_attrs = { %$attrs };
1466 delete $attrs->{$_} for qw/where bind rows offset/;
1467 delete $sub_attrs->{$_} for qw/for collapse select order_by/;
1469 my $alias = $attrs->{alias};
1471 # create subquery select list
1472 my $sub_select = [ grep { $_ =~ /^$alias\./ } @{$attrs->{select}} ];
1474 # bring over all non-collapse-induced order_by into the inner query (if any)
1475 # the outer one will have to keep them all
1476 if (my $ord_cnt = @{$attrs->{order_by}} - @{$attrs->{_collapse_order_by}} ) {
1477 $sub_attrs->{order_by} = [
1478 @{$attrs->{order_by}}[ 0 .. ($#{$attrs->{order_by}} - $ord_cnt - 1) ]
1483 # mangle the head of the {from}
1484 my $self_ident = shift @$from;
1486 my %join_info = map { $_->[0]{-alias} => $_->[0] } (@$from);
1490 # decide which parts of the join will remain on the inside
1492 # this is not a very viable optimisation, but it was written
1493 # before I realised this, so might as well remain. We can throw
1494 # away _any_ branches of the join tree that are:
1495 # 1) not mentioned in the condition/order
1496 # 2) left-join leaves (or left-join leaf chains)
1497 # Most of the join ocnditions will not satisfy this, but for real
1498 # complex queries some might, and we might make some RDBMS happy.
1501 # since we do not have introspectable SQLA, we fall back to ugly
1502 # scanning of raw SQL for WHERE, and for pieces of ORDER BY
1503 # in order to determine what goes into %inner_joins
1504 # It may not be very efficient, but it's a reasonable stop-gap
1506 # produce stuff unquoted, so it can be scanned
1507 my $sql_maker = $self->sql_maker;
1508 local $sql_maker->{quote_char};
1511 { ref $_ ? $_->[0] : $_ }
1512 $sql_maker->_order_by_chunks ($sub_attrs->{order_by})
1515 my $where_sql = $sql_maker->where ($where);
1518 for my $alias (keys %join_info) {
1520 # any table alias found on a column name in where or order_by
1521 # gets included in %inner_joins
1522 # Also any parent joins that are needed to reach this particular alias
1523 for my $piece ($where_sql, @order_by ) {
1524 if ($piece =~ /\b$alias\./) {
1525 $inner_joins{$alias} = 1;
1531 # scan for non-leaf/non-left joins and mark as needed
1532 # also mark all ancestor joins that are needed to reach this particular alias
1533 # (e.g. join => { cds => 'tracks' } - tracks will bring cds too )
1535 # traverse by the size of the -join_path i.e. reverse depth first
1536 for my $alias (sort { @{$join_info{$b}{-join_path}} <=> @{$join_info{$a}{-join_path}} } (keys %join_info) ) {
1538 my $j = $join_info{$alias};
1539 $inner_joins{$alias} = 1 if (! $j->{-join_type} || ($j->{-join_type} !~ /^left$/i) );
1541 if ($inner_joins{$alias}) {
1542 $inner_joins{$_} = 1 for (@{$j->{-join_path}});
1546 # construct the inner $from for the subquery
1547 my $inner_from = [ $self_ident ];
1548 if (keys %inner_joins) {
1549 for my $j (@$from) {
1550 push @$inner_from, $j if $inner_joins{$j->[0]{-alias}};
1553 # if a multi-type join was needed in the subquery ("multi" is indicated by
1554 # presence in {collapse}) - add a group_by to simulate the collapse in the subq
1555 for my $alias (keys %inner_joins) {
1557 # the dot comes from some weirdness in collapse
1558 # remove after the rewrite
1559 if ($attrs->{collapse}{".$alias"}) {
1560 $sub_attrs->{group_by} = $sub_select;
1566 # generate the subquery
1567 my $subq = $self->_select_args_to_query (
1574 # put it back in $from
1575 unshift @$from, { $alias => $subq };
1577 # This is totally horrific - the $where ends up in both the inner and outer query
1578 # Unfortunately not much can be done until SQLA2 introspection arrives
1580 # OTOH it can be seen as a plus: <ash> (notes that this query would make a DBA cry ;)
1581 return ($from, $select, $where, $attrs);
1584 sub _resolve_ident_sources {
1585 my ($self, $ident) = @_;
1587 my $alias2source = {};
1589 # the reason this is so contrived is that $ident may be a {from}
1590 # structure, specifying multiple tables to join
1591 if ( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1592 # this is compat mode for insert/update/delete which do not deal with aliases
1593 $alias2source->{me} = $ident;
1595 elsif (ref $ident eq 'ARRAY') {
1599 if (ref $_ eq 'HASH') {
1602 if (ref $_ eq 'ARRAY' and ref $_->[0] eq 'HASH') {
1606 $alias2source->{$tabinfo->{-alias}} = $tabinfo->{-source_handle}->resolve
1607 if ($tabinfo->{-source_handle});
1611 return $alias2source;
1615 my ($self, $source, $attrs) = @_;
1617 my $tmp_attrs = { %$attrs };
1619 # take off any limits, record_filter is cdbi, and no point of ordering a count
1620 delete $tmp_attrs->{$_} for (qw/select as rows offset order_by record_filter/);
1622 # overwrite the selector
1623 $tmp_attrs->{select} = { count => '*' };
1625 my $tmp_rs = $source->resultset_class->new($source, $tmp_attrs);
1626 my ($count) = $tmp_rs->cursor->next;
1628 # if the offset/rows attributes are still present, we did not use
1629 # a subquery, so we need to make the calculations in software
1630 $count -= $attrs->{offset} if $attrs->{offset};
1631 $count = $attrs->{rows} if $attrs->{rows} and $attrs->{rows} < $count;
1632 $count = 0 if ($count < 0);
1638 my ($self, $source, $attrs) = @_;
1640 # copy for the subquery, we need to do some adjustments to it too
1641 my $sub_attrs = { %$attrs };
1643 # these can not go in the subquery, and there is no point of ordering it
1644 delete $sub_attrs->{$_} for qw/collapse select as order_by/;
1646 # if we prefetch, we group_by primary keys only as this is what we would get out of the rs via ->next/->all
1647 # simply deleting group_by suffices, as the code below will re-fill it
1648 # Note: we check $attrs, as $sub_attrs has collapse deleted
1649 if (ref $attrs->{collapse} and keys %{$attrs->{collapse}} ) {
1650 delete $sub_attrs->{group_by};
1653 $sub_attrs->{group_by} ||= [ map { "$attrs->{alias}.$_" } ($source->primary_columns) ];
1654 $sub_attrs->{select} = $self->_grouped_count_select ($source, $sub_attrs);
1657 count_subq => $source->resultset_class->new ($source, $sub_attrs )->as_query
1660 # the subquery replaces this
1661 delete $attrs->{$_} for qw/where bind collapse group_by having having_bind rows offset/;
1663 return $self->count ($source, $attrs);
1667 # Returns a SELECT to go with a supplied GROUP BY
1668 # (caled by count_grouped so a group_by is present)
1669 # Most databases expect them to match, but some
1670 # choke in various ways.
1672 sub _grouped_count_select {
1673 my ($self, $source, $rs_args) = @_;
1674 return $rs_args->{group_by};
1677 sub source_bind_attributes {
1678 my ($self, $source) = @_;
1680 my $bind_attributes;
1681 foreach my $column ($source->columns) {
1683 my $data_type = $source->column_info($column)->{data_type} || '';
1684 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
1688 return $bind_attributes;
1695 =item Arguments: $ident, $select, $condition, $attrs
1699 Handle a SQL select statement.
1705 my ($ident, $select, $condition, $attrs) = @_;
1706 return $self->cursor_class->new($self, \@_, $attrs);
1711 my ($rv, $sth, @bind) = $self->_select(@_);
1712 my @row = $sth->fetchrow_array;
1713 my @nextrow = $sth->fetchrow_array if @row;
1714 if(@row && @nextrow) {
1715 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
1717 # Need to call finish() to work round broken DBDs
1726 =item Arguments: $sql
1730 Returns a L<DBI> sth (statement handle) for the supplied SQL.
1735 my ($self, $dbh, $sql) = @_;
1737 # 3 is the if_active parameter which avoids active sth re-use
1738 my $sth = $self->disable_sth_caching
1739 ? $dbh->prepare($sql)
1740 : $dbh->prepare_cached($sql, {}, 3);
1742 # XXX You would think RaiseError would make this impossible,
1743 # but apparently that's not true :(
1744 $self->throw_exception($dbh->errstr) if !$sth;
1750 my ($self, $sql) = @_;
1751 $self->dbh_do('_dbh_sth', $sql);
1754 sub _dbh_columns_info_for {
1755 my ($self, $dbh, $table) = @_;
1757 if ($dbh->can('column_info')) {
1760 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
1761 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
1763 while ( my $info = $sth->fetchrow_hashref() ){
1765 $column_info{data_type} = $info->{TYPE_NAME};
1766 $column_info{size} = $info->{COLUMN_SIZE};
1767 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
1768 $column_info{default_value} = $info->{COLUMN_DEF};
1769 my $col_name = $info->{COLUMN_NAME};
1770 $col_name =~ s/^\"(.*)\"$/$1/;
1772 $result{$col_name} = \%column_info;
1775 return \%result if !$@ && scalar keys %result;
1779 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
1781 my @columns = @{$sth->{NAME_lc}};
1782 for my $i ( 0 .. $#columns ){
1784 $column_info{data_type} = $sth->{TYPE}->[$i];
1785 $column_info{size} = $sth->{PRECISION}->[$i];
1786 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
1788 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
1789 $column_info{data_type} = $1;
1790 $column_info{size} = $2;
1793 $result{$columns[$i]} = \%column_info;
1797 foreach my $col (keys %result) {
1798 my $colinfo = $result{$col};
1799 my $type_num = $colinfo->{data_type};
1801 if(defined $type_num && $dbh->can('type_info')) {
1802 my $type_info = $dbh->type_info($type_num);
1803 $type_name = $type_info->{TYPE_NAME} if $type_info;
1804 $colinfo->{data_type} = $type_name if $type_name;
1811 sub columns_info_for {
1812 my ($self, $table) = @_;
1813 $self->dbh_do('_dbh_columns_info_for', $table);
1816 =head2 last_insert_id
1818 Return the row id of the last insert.
1822 sub _dbh_last_insert_id {
1823 # All Storage's need to register their own _dbh_last_insert_id
1824 # the old SQLite-based method was highly inappropriate
1827 my $class = ref $self;
1828 $self->throw_exception (<<EOE);
1830 No _dbh_last_insert_id() method found in $class.
1831 Since the method of obtaining the autoincrement id of the last insert
1832 operation varies greatly between different databases, this method must be
1833 individually implemented for every storage class.
1837 sub last_insert_id {
1839 $self->dbh_do('_dbh_last_insert_id', @_);
1844 Returns the database driver name.
1848 sub sqlt_type { shift->dbh->{Driver}->{Name} }
1850 =head2 bind_attribute_by_data_type
1852 Given a datatype from column info, returns a database specific bind
1853 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
1854 let the database planner just handle it.
1856 Generally only needed for special case column types, like bytea in postgres.
1860 sub bind_attribute_by_data_type {
1864 =head2 is_datatype_numeric
1866 Given a datatype from column_info, returns a boolean value indicating if
1867 the current RDBMS considers it a numeric value. This controls how
1868 L<DBIx::Class::Row/set_column> decides whether to mark the column as
1869 dirty - when the datatype is deemed numeric a C<< != >> comparison will
1870 be performed instead of the usual C<eq>.
1874 sub is_datatype_numeric {
1875 my ($self, $dt) = @_;
1877 return 0 unless $dt;
1879 return $dt =~ /^ (?:
1880 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
1885 =head2 create_ddl_dir (EXPERIMENTAL)
1889 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
1893 Creates a SQL file based on the Schema, for each of the specified
1894 database engines in C<\@databases> in the given directory.
1895 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
1897 Given a previous version number, this will also create a file containing
1898 the ALTER TABLE statements to transform the previous schema into the
1899 current one. Note that these statements may contain C<DROP TABLE> or
1900 C<DROP COLUMN> statements that can potentially destroy data.
1902 The file names are created using the C<ddl_filename> method below, please
1903 override this method in your schema if you would like a different file
1904 name format. For the ALTER file, the same format is used, replacing
1905 $version in the name with "$preversion-$version".
1907 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
1908 The most common value for this would be C<< { add_drop_table => 1 } >>
1909 to have the SQL produced include a C<DROP TABLE> statement for each table
1910 created. For quoting purposes supply C<quote_table_names> and
1911 C<quote_field_names>.
1913 If no arguments are passed, then the following default values are assumed:
1917 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
1919 =item version - $schema->schema_version
1921 =item directory - './'
1923 =item preversion - <none>
1927 By default, C<\%sqlt_args> will have
1929 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
1931 merged with the hash passed in. To disable any of those features, pass in a
1932 hashref like the following
1934 { ignore_constraint_names => 0, # ... other options }
1937 Note that this feature is currently EXPERIMENTAL and may not work correctly
1938 across all databases, or fully handle complex relationships.
1940 WARNING: Please check all SQL files created, before applying them.
1944 sub create_ddl_dir {
1945 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
1947 if(!$dir || !-d $dir) {
1948 carp "No directory given, using ./\n";
1951 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
1952 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
1954 my $schema_version = $schema->schema_version || '1.x';
1955 $version ||= $schema_version;
1958 add_drop_table => 1,
1959 ignore_constraint_names => 1,
1960 ignore_index_names => 1,
1964 $self->throw_exception(q{Can't create a ddl file without SQL::Translator 0.09003: '}
1965 . $self->_check_sqlt_message . q{'})
1966 if !$self->_check_sqlt_version;
1968 my $sqlt = SQL::Translator->new( $sqltargs );
1970 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
1971 my $sqlt_schema = $sqlt->translate({ data => $schema })
1972 or $self->throw_exception ($sqlt->error);
1974 foreach my $db (@$databases) {
1976 $sqlt->{schema} = $sqlt_schema;
1977 $sqlt->producer($db);
1980 my $filename = $schema->ddl_filename($db, $version, $dir);
1981 if (-e $filename && ($version eq $schema_version )) {
1982 # if we are dumping the current version, overwrite the DDL
1983 carp "Overwriting existing DDL file - $filename";
1987 my $output = $sqlt->translate;
1989 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
1992 if(!open($file, ">$filename")) {
1993 $self->throw_exception("Can't open $filename for writing ($!)");
1996 print $file $output;
1999 next unless ($preversion);
2001 require SQL::Translator::Diff;
2003 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2004 if(!-e $prefilename) {
2005 carp("No previous schema file found ($prefilename)");
2009 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2011 carp("Overwriting existing diff file - $difffile");
2017 my $t = SQL::Translator->new($sqltargs);
2022 or $self->throw_exception ($t->error);
2024 my $out = $t->translate( $prefilename )
2025 or $self->throw_exception ($t->error);
2027 $source_schema = $t->schema;
2029 $source_schema->name( $prefilename )
2030 unless ( $source_schema->name );
2033 # The "new" style of producers have sane normalization and can support
2034 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2035 # And we have to diff parsed SQL against parsed SQL.
2036 my $dest_schema = $sqlt_schema;
2038 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2039 my $t = SQL::Translator->new($sqltargs);
2044 or $self->throw_exception ($t->error);
2046 my $out = $t->translate( $filename )
2047 or $self->throw_exception ($t->error);
2049 $dest_schema = $t->schema;
2051 $dest_schema->name( $filename )
2052 unless $dest_schema->name;
2055 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2059 if(!open $file, ">$difffile") {
2060 $self->throw_exception("Can't write to $difffile ($!)");
2068 =head2 deployment_statements
2072 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2076 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2078 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2079 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2081 C<$directory> is used to return statements from files in a previously created
2082 L</create_ddl_dir> directory and is optional. The filenames are constructed
2083 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2085 If no C<$directory> is specified then the statements are constructed on the
2086 fly using L<SQL::Translator> and C<$version> is ignored.
2088 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2092 sub deployment_statements {
2093 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2094 # Need to be connected to get the correct sqlt_type
2095 $self->ensure_connected() unless $type;
2096 $type ||= $self->sqlt_type;
2097 $version ||= $schema->schema_version || '1.x';
2099 my $filename = $schema->ddl_filename($type, $version, $dir);
2103 open($file, "<$filename")
2104 or $self->throw_exception("Can't open $filename ($!)");
2107 return join('', @rows);
2110 $self->throw_exception(q{Can't deploy without SQL::Translator 0.09003: '}
2111 . $self->_check_sqlt_message . q{'})
2112 if !$self->_check_sqlt_version;
2114 require SQL::Translator::Parser::DBIx::Class;
2115 eval qq{use SQL::Translator::Producer::${type}};
2116 $self->throw_exception($@) if $@;
2118 # sources needs to be a parser arg, but for simplicty allow at top level
2120 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2121 if exists $sqltargs->{sources};
2123 my $tr = SQL::Translator->new(%$sqltargs);
2124 SQL::Translator::Parser::DBIx::Class::parse( $tr, $schema );
2125 return "SQL::Translator::Producer::${type}"->can('produce')->($tr);
2129 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2132 return if($line =~ /^--/);
2134 # next if($line =~ /^DROP/m);
2135 return if($line =~ /^BEGIN TRANSACTION/m);
2136 return if($line =~ /^COMMIT/m);
2137 return if $line =~ /^\s+$/; # skip whitespace only
2138 $self->_query_start($line);
2140 $self->dbh->do($line); # shouldn't be using ->dbh ?
2143 carp qq{$@ (running "${line}")};
2145 $self->_query_end($line);
2147 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2148 if (@statements > 1) {
2149 foreach my $statement (@statements) {
2150 $deploy->( $statement );
2153 elsif (@statements == 1) {
2154 foreach my $line ( split(";\n", $statements[0])) {
2160 =head2 datetime_parser
2162 Returns the datetime parser class
2166 sub datetime_parser {
2168 return $self->{datetime_parser} ||= do {
2169 $self->ensure_connected;
2170 $self->build_datetime_parser(@_);
2174 =head2 datetime_parser_type
2176 Defines (returns) the datetime parser class - currently hardwired to
2177 L<DateTime::Format::MySQL>
2181 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2183 =head2 build_datetime_parser
2185 See L</datetime_parser>
2189 sub build_datetime_parser {
2191 my $type = $self->datetime_parser_type(@_);
2193 $self->throw_exception("Couldn't load ${type}: $@") if $@;
2198 my $_check_sqlt_version; # private
2199 my $_check_sqlt_message; # private
2200 sub _check_sqlt_version {
2201 return $_check_sqlt_version if defined $_check_sqlt_version;
2202 eval 'use SQL::Translator "0.09003"';
2203 $_check_sqlt_message = $@ || '';
2204 $_check_sqlt_version = !$@;
2207 sub _check_sqlt_message {
2208 _check_sqlt_version if !defined $_check_sqlt_message;
2209 $_check_sqlt_message;
2213 =head2 is_replicating
2215 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2216 replicate from a master database. Default is undef, which is the result
2217 returned by databases that don't support replication.
2221 sub is_replicating {
2226 =head2 lag_behind_master
2228 Returns a number that represents a certain amount of lag behind a master db
2229 when a given storage is replicating. The number is database dependent, but
2230 starts at zero and increases with the amount of lag. Default in undef
2234 sub lag_behind_master {
2240 return if !$self->_dbh;
2249 =head2 DBIx::Class and AutoCommit
2251 DBIx::Class can do some wonderful magic with handling exceptions,
2252 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2253 combined with C<txn_do> for transaction support.
2255 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2256 in an assumed transaction between commits, and you're telling us you'd
2257 like to manage that manually. A lot of the magic protections offered by
2258 this module will go away. We can't protect you from exceptions due to database
2259 disconnects because we don't know anything about how to restart your
2260 transactions. You're on your own for handling all sorts of exceptional
2261 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2268 Matt S. Trout <mst@shadowcatsystems.co.uk>
2270 Andy Grundman <andy@hybridized.org>
2274 You may distribute this code under the same terms as Perl itself.