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::SQLAHacks;
11 use DBIx::Class::Storage::DBI::Cursor;
12 use DBIx::Class::Storage::Statistics;
13 use Scalar::Util qw/blessed weaken/;
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 savepoints/
20 # the values for these accessors are picked out (and deleted) from
21 # the attribute hashref passed to connect_info
22 my @storage_options = qw/
23 on_connect_do on_disconnect_do disable_sth_caching unsafe auto_savepoint
25 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
28 # default cursor class, overridable in connect_info attributes
29 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
31 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class/);
32 __PACKAGE__->sql_maker_class('DBIx::Class::SQLAHacks');
37 DBIx::Class::Storage::DBI - DBI storage handler
41 my $schema = MySchema->connect('dbi:SQLite:my.db');
43 $schema->storage->debug(1);
44 $schema->dbh_do("DROP TABLE authors");
46 $schema->resultset('Book')->search({
47 written_on => $schema->storage->datetime_parser(DateTime->now)
52 This class represents the connection to an RDBMS via L<DBI>. See
53 L<DBIx::Class::Storage> for general information. This pod only
54 documents DBI-specific methods and behaviors.
61 my $new = shift->next::method(@_);
63 $new->transaction_depth(0);
64 $new->_sql_maker_opts({});
65 $new->{savepoints} = [];
66 $new->{_in_dbh_do} = 0;
74 This method is normally called by L<DBIx::Class::Schema/connection>, which
75 encapsulates its argument list in an arrayref before passing them here.
77 The argument list may contain:
83 The same 4-element argument set one would normally pass to
84 L<DBI/connect>, optionally followed by
85 L<extra attributes|/DBIx::Class specific connection attributes>
86 recognized by DBIx::Class:
88 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
92 A single code reference which returns a connected
93 L<DBI database handle|DBI/connect> optionally followed by
94 L<extra attributes|/DBIx::Class specific connection attributes> recognized
97 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
101 A single hashref with all the attributes and the dsn/user/password
104 $connect_info_args = [{
112 This is particularly useful for L<Catalyst> based applications, allowing the
113 following config (L<Config::General> style):
118 dsn dbi:mysql:database=test
127 Please note that the L<DBI> docs recommend that you always explicitly
128 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
129 recommends that it be set to I<1>, and that you perform transactions
130 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
131 to I<1> if you do not do explicitly set it to zero. This is the default
132 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
134 =head3 DBIx::Class specific connection attributes
136 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
137 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
138 the following connection options. These options can be mixed in with your other
139 L<DBI> connection attributes, or placed in a seperate hashref
140 (C<\%extra_attributes>) as shown above.
142 Every time C<connect_info> is invoked, any previous settings for
143 these options will be cleared before setting the new ones, regardless of
144 whether any options are specified in the new C<connect_info>.
151 Specifies things to do immediately after connecting or re-connecting to
152 the database. Its value may contain:
158 This contains one SQL statement to execute.
160 =item an array reference
162 This contains SQL statements to execute in order. Each element contains
163 a string or a code reference that returns a string.
165 =item a code reference
167 This contains some code to execute. Unlike code references within an
168 array reference, its return value is ignored.
172 =item on_disconnect_do
174 Takes arguments in the same form as L</on_connect_do> and executes them
175 immediately before disconnecting from the database.
177 Note, this only runs if you explicitly call L</disconnect> on the
180 =item disable_sth_caching
182 If set to a true value, this option will disable the caching of
183 statement handles via L<DBI/prepare_cached>.
187 Sets the limit dialect. This is useful for JDBC-bridge among others
188 where the remote SQL-dialect cannot be determined by the name of the
189 driver alone. See also L<SQL::Abstract::Limit>.
193 Specifies what characters to use to quote table and column names. If
194 you use this you will want to specify L</name_sep> as well.
196 C<quote_char> expects either a single character, in which case is it
197 is placed on either side of the table/column name, or an arrayref of length
198 2 in which case the table/column name is placed between the elements.
200 For example under MySQL you should use C<< quote_char => '`' >>, and for
201 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
205 This only needs to be used in conjunction with C<quote_char>, and is used to
206 specify the charecter that seperates elements (schemas, tables, columns) from
207 each other. In most cases this is simply a C<.>.
209 The consequences of not supplying this value is that L<SQL::Abstract>
210 will assume DBIx::Class' uses of aliases to be complete column
211 names. The output will look like I<"me.name"> when it should actually
216 This Storage driver normally installs its own C<HandleError>, sets
217 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
218 all database handles, including those supplied by a coderef. It does this
219 so that it can have consistent and useful error behavior.
221 If you set this option to a true value, Storage will not do its usual
222 modifications to the database handle's attributes, and instead relies on
223 the settings in your connect_info DBI options (or the values you set in
224 your connection coderef, in the case that you are connecting via coderef).
226 Note that your custom settings can cause Storage to malfunction,
227 especially if you set a C<HandleError> handler that suppresses exceptions
228 and/or disable C<RaiseError>.
232 If this option is true, L<DBIx::Class> will use savepoints when nesting
233 transactions, making it possible to recover from failure in the inner
234 transaction without having to abort all outer transactions.
238 Use this argument to supply a cursor class other than the default
239 L<DBIx::Class::Storage::DBI::Cursor>.
243 Some real-life examples of arguments to L</connect_info> and
244 L<DBIx::Class::Schema/connect>
246 # Simple SQLite connection
247 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
250 ->connect_info([ sub { DBI->connect(...) } ]);
252 # A bit more complicated
259 { quote_char => q{"}, name_sep => q{.} },
263 # Equivalent to the previous example
269 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
273 # Same, but with hashref as argument
274 # See parse_connect_info for explanation
277 dsn => 'dbi:Pg:dbname=foo',
279 password => 'my_pg_password',
286 # Subref + DBIx::Class-specific connection options
289 sub { DBI->connect(...) },
293 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
294 disable_sth_caching => 1,
304 my ($self, $info_arg) = @_;
306 return $self->_connect_info if !$info_arg;
308 my @args = @$info_arg; # take a shallow copy for further mutilation
309 $self->_connect_info([@args]); # copy for _connect_info
312 # combine/pre-parse arguments depending on invocation style
315 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
316 %attrs = %{ $args[1] || {} };
319 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
320 %attrs = %{$args[0]};
322 for (qw/password user dsn/) {
323 unshift @args, delete $attrs{$_};
326 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
328 % { $args[3] || {} },
329 % { $args[4] || {} },
331 @args = @args[0,1,2];
334 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
335 # the new set of options
336 $self->_sql_maker(undef);
337 $self->_sql_maker_opts({});
340 for my $storage_opt (@storage_options, 'cursor_class') { # @storage_options is declared at the top of the module
341 if(my $value = delete $attrs{$storage_opt}) {
342 $self->$storage_opt($value);
345 for my $sql_maker_opt (qw/limit_dialect quote_char name_sep/) {
346 if(my $opt_val = delete $attrs{$sql_maker_opt}) {
347 $self->_sql_maker_opts->{$sql_maker_opt} = $opt_val;
352 %attrs = () if (ref $args[0] eq 'CODE'); # _connect() never looks past $args[0] in this case
354 $self->_dbi_connect_info([@args, keys %attrs ? \%attrs : ()]);
355 $self->_connect_info;
360 This method is deprecated in favour of setting via L</connect_info>.
365 Arguments: ($subref | $method_name), @extra_coderef_args?
367 Execute the given $subref or $method_name using the new exception-based
368 connection management.
370 The first two arguments will be the storage object that C<dbh_do> was called
371 on and a database handle to use. Any additional arguments will be passed
372 verbatim to the called subref as arguments 2 and onwards.
374 Using this (instead of $self->_dbh or $self->dbh) ensures correct
375 exception handling and reconnection (or failover in future subclasses).
377 Your subref should have no side-effects outside of the database, as
378 there is the potential for your subref to be partially double-executed
379 if the database connection was stale/dysfunctional.
383 my @stuff = $schema->storage->dbh_do(
385 my ($storage, $dbh, @cols) = @_;
386 my $cols = join(q{, }, @cols);
387 $dbh->selectrow_array("SELECT $cols FROM foo");
398 my $dbh = $self->_dbh;
400 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
401 || $self->{transaction_depth};
403 local $self->{_in_dbh_do} = 1;
406 my $want_array = wantarray;
409 $self->_verify_pid if $dbh;
411 $self->_populate_dbh;
416 @result = $self->$code($dbh, @_);
418 elsif(defined $want_array) {
419 $result[0] = $self->$code($dbh, @_);
422 $self->$code($dbh, @_);
427 if(!$exception) { return $want_array ? @result : $result[0] }
429 $self->throw_exception($exception) if $self->connected;
431 # We were not connected - reconnect and retry, but let any
432 # exception fall right through this time
433 $self->_populate_dbh;
434 $self->$code($self->_dbh, @_);
437 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
438 # It also informs dbh_do to bypass itself while under the direction of txn_do,
439 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
444 ref $coderef eq 'CODE' or $self->throw_exception
445 ('$coderef must be a CODE reference');
447 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
449 local $self->{_in_dbh_do} = 1;
452 my $want_array = wantarray;
457 $self->_verify_pid if $self->_dbh;
458 $self->_populate_dbh if !$self->_dbh;
462 @result = $coderef->(@_);
464 elsif(defined $want_array) {
465 $result[0] = $coderef->(@_);
474 if(!$exception) { return $want_array ? @result : $result[0] }
476 if($tried++ > 0 || $self->connected) {
477 eval { $self->txn_rollback };
478 my $rollback_exception = $@;
479 if($rollback_exception) {
480 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
481 $self->throw_exception($exception) # propagate nested rollback
482 if $rollback_exception =~ /$exception_class/;
484 $self->throw_exception(
485 "Transaction aborted: ${exception}. "
486 . "Rollback failed: ${rollback_exception}"
489 $self->throw_exception($exception)
492 # We were not connected, and was first try - reconnect and retry
494 $self->_populate_dbh;
500 Our C<disconnect> method also performs a rollback first if the
501 database is not in C<AutoCommit> mode.
508 if( $self->connected ) {
509 my $connection_do = $self->on_disconnect_do;
510 $self->_do_connection_actions($connection_do) if ref($connection_do);
512 $self->_dbh->rollback unless $self->_dbh_autocommit;
513 $self->_dbh->disconnect;
519 =head2 with_deferred_fk_checks
523 =item Arguments: C<$coderef>
525 =item Return Value: The return value of $coderef
529 Storage specific method to run the code ref with FK checks deferred or
530 in MySQL's case disabled entirely.
534 # Storage subclasses should override this
535 sub with_deferred_fk_checks {
536 my ($self, $sub) = @_;
544 if(my $dbh = $self->_dbh) {
545 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
552 return 0 if !$self->_dbh;
554 return ($dbh->FETCH('Active') && $dbh->ping);
560 # handle pid changes correctly
561 # NOTE: assumes $self->_dbh is a valid $dbh
565 return if defined $self->_conn_pid && $self->_conn_pid == $$;
567 $self->_dbh->{InactiveDestroy} = 1;
574 sub ensure_connected {
577 unless ($self->connected) {
578 $self->_populate_dbh;
584 Returns the dbh - a data base handle of class L<DBI>.
591 $self->ensure_connected;
595 sub _sql_maker_args {
598 return ( bindtype=>'columns', array_datatypes => 1, limit_dialect => $self->dbh, %{$self->_sql_maker_opts} );
603 unless ($self->_sql_maker) {
604 my $sql_maker_class = $self->sql_maker_class;
605 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
607 return $self->_sql_maker;
614 my @info = @{$self->_dbi_connect_info || []};
615 $self->_dbh($self->_connect(@info));
617 # Always set the transaction depth on connect, since
618 # there is no transaction in progress by definition
619 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
621 if(ref $self eq 'DBIx::Class::Storage::DBI') {
622 my $driver = $self->_dbh->{Driver}->{Name};
623 if ($self->load_optional_class("DBIx::Class::Storage::DBI::${driver}")) {
624 bless $self, "DBIx::Class::Storage::DBI::${driver}";
629 $self->_conn_pid($$);
630 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
632 my $connection_do = $self->on_connect_do;
633 $self->_do_connection_actions($connection_do) if $connection_do;
636 sub _do_connection_actions {
638 my $connection_do = shift;
640 if (!ref $connection_do) {
641 $self->_do_query($connection_do);
643 elsif (ref $connection_do eq 'ARRAY') {
644 $self->_do_query($_) foreach @$connection_do;
646 elsif (ref $connection_do eq 'CODE') {
647 $connection_do->($self);
650 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref $connection_do) );
657 my ($self, $action) = @_;
659 if (ref $action eq 'CODE') {
660 $action = $action->($self);
661 $self->_do_query($_) foreach @$action;
664 # Most debuggers expect ($sql, @bind), so we need to exclude
665 # the attribute hash which is the second argument to $dbh->do
666 # furthermore the bind values are usually to be presented
667 # as named arrayref pairs, so wrap those here too
668 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
669 my $sql = shift @do_args;
670 my $attrs = shift @do_args;
671 my @bind = map { [ undef, $_ ] } @do_args;
673 $self->_query_start($sql, @bind);
674 $self->_dbh->do($sql, $attrs, @do_args);
675 $self->_query_end($sql, @bind);
682 my ($self, @info) = @_;
684 $self->throw_exception("You failed to provide any connection info")
687 my ($old_connect_via, $dbh);
689 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
690 $old_connect_via = $DBI::connect_via;
691 $DBI::connect_via = 'connect';
695 if(ref $info[0] eq 'CODE') {
699 $dbh = DBI->connect(@info);
702 if($dbh && !$self->unsafe) {
703 my $weak_self = $self;
705 $dbh->{HandleError} = sub {
707 $weak_self->throw_exception("DBI Exception: $_[0]");
710 croak ("DBI Exception: $_[0]");
713 $dbh->{ShowErrorStatement} = 1;
714 $dbh->{RaiseError} = 1;
715 $dbh->{PrintError} = 0;
719 $DBI::connect_via = $old_connect_via if $old_connect_via;
721 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
724 $self->_dbh_autocommit($dbh->{AutoCommit});
730 my ($self, $name) = @_;
732 $name = $self->_svp_generate_name
733 unless defined $name;
735 $self->throw_exception ("You can't use savepoints outside a transaction")
736 if $self->{transaction_depth} == 0;
738 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
739 unless $self->can('_svp_begin');
741 push @{ $self->{savepoints} }, $name;
743 $self->debugobj->svp_begin($name) if $self->debug;
745 return $self->_svp_begin($name);
749 my ($self, $name) = @_;
751 $self->throw_exception ("You can't use savepoints outside a transaction")
752 if $self->{transaction_depth} == 0;
754 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
755 unless $self->can('_svp_release');
758 $self->throw_exception ("Savepoint '$name' does not exist")
759 unless grep { $_ eq $name } @{ $self->{savepoints} };
761 # Dig through the stack until we find the one we are releasing. This keeps
762 # the stack up to date.
765 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
767 $name = pop @{ $self->{savepoints} };
770 $self->debugobj->svp_release($name) if $self->debug;
772 return $self->_svp_release($name);
776 my ($self, $name) = @_;
778 $self->throw_exception ("You can't use savepoints outside a transaction")
779 if $self->{transaction_depth} == 0;
781 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
782 unless $self->can('_svp_rollback');
785 # If they passed us a name, verify that it exists in the stack
786 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
787 $self->throw_exception("Savepoint '$name' does not exist!");
790 # Dig through the stack until we find the one we are releasing. This keeps
791 # the stack up to date.
792 while(my $s = pop(@{ $self->{savepoints} })) {
793 last if($s eq $name);
795 # Add the savepoint back to the stack, as a rollback doesn't remove the
796 # named savepoint, only everything after it.
797 push(@{ $self->{savepoints} }, $name);
799 # We'll assume they want to rollback to the last savepoint
800 $name = $self->{savepoints}->[-1];
803 $self->debugobj->svp_rollback($name) if $self->debug;
805 return $self->_svp_rollback($name);
808 sub _svp_generate_name {
811 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
816 $self->ensure_connected();
817 if($self->{transaction_depth} == 0) {
818 $self->debugobj->txn_begin()
820 # this isn't ->_dbh-> because
821 # we should reconnect on begin_work
822 # for AutoCommit users
823 $self->dbh->begin_work;
824 } elsif ($self->auto_savepoint) {
827 $self->{transaction_depth}++;
832 if ($self->{transaction_depth} == 1) {
833 my $dbh = $self->_dbh;
834 $self->debugobj->txn_commit()
837 $self->{transaction_depth} = 0
838 if $self->_dbh_autocommit;
840 elsif($self->{transaction_depth} > 1) {
841 $self->{transaction_depth}--;
843 if $self->auto_savepoint;
849 my $dbh = $self->_dbh;
851 if ($self->{transaction_depth} == 1) {
852 $self->debugobj->txn_rollback()
854 $self->{transaction_depth} = 0
855 if $self->_dbh_autocommit;
858 elsif($self->{transaction_depth} > 1) {
859 $self->{transaction_depth}--;
860 if ($self->auto_savepoint) {
866 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
871 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
872 $error =~ /$exception_class/ and $self->throw_exception($error);
873 # ensure that a failed rollback resets the transaction depth
874 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
875 $self->throw_exception($error);
879 # This used to be the top-half of _execute. It was split out to make it
880 # easier to override in NoBindVars without duping the rest. It takes up
881 # all of _execute's args, and emits $sql, @bind.
882 sub _prep_for_execute {
883 my ($self, $op, $extra_bind, $ident, $args) = @_;
885 if( blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
886 $ident = $ident->from();
889 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
892 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
894 return ($sql, \@bind);
897 sub _fix_bind_params {
898 my ($self, @bind) = @_;
900 ### Turn @bind from something like this:
901 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
903 ### ( "'1'", "'1'", "'3'" )
906 if ( defined( $_ && $_->[1] ) ) {
907 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
914 my ( $self, $sql, @bind ) = @_;
916 if ( $self->debug ) {
917 @bind = $self->_fix_bind_params(@bind);
919 $self->debugobj->query_start( $sql, @bind );
924 my ( $self, $sql, @bind ) = @_;
926 if ( $self->debug ) {
927 @bind = $self->_fix_bind_params(@bind);
928 $self->debugobj->query_end( $sql, @bind );
933 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
935 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
937 $self->_query_start( $sql, @$bind );
939 my $sth = $self->sth($sql,$op);
941 my $placeholder_index = 1;
943 foreach my $bound (@$bind) {
945 my($column_name, @data) = @$bound;
947 if ($bind_attributes) {
948 $attributes = $bind_attributes->{$column_name}
949 if defined $bind_attributes->{$column_name};
952 foreach my $data (@data) {
954 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
956 $sth->bind_param($placeholder_index, $data, $attributes);
957 $placeholder_index++;
961 # Can this fail without throwing an exception anyways???
962 my $rv = $sth->execute();
963 $self->throw_exception($sth->errstr) if !$rv;
965 $self->_query_end( $sql, @$bind );
967 return (wantarray ? ($rv, $sth, @$bind) : $rv);
972 $self->dbh_do('_dbh_execute', @_)
976 my ($self, $source, $to_insert) = @_;
978 my $ident = $source->from;
979 my $bind_attributes = $self->source_bind_attributes($source);
981 my $updated_cols = {};
983 $self->ensure_connected;
984 foreach my $col ( $source->columns ) {
985 if ( !defined $to_insert->{$col} ) {
986 my $col_info = $source->column_info($col);
988 if ( $col_info->{auto_nextval} ) {
989 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch( 'nextval', $col_info->{sequence} || $self->_dbh_get_autoinc_seq($self->dbh, $source) );
994 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
996 return $updated_cols;
999 ## Still not quite perfect, and EXPERIMENTAL
1000 ## Currently it is assumed that all values passed will be "normal", i.e. not
1001 ## scalar refs, or at least, all the same type as the first set, the statement is
1002 ## only prepped once.
1004 my ($self, $source, $cols, $data) = @_;
1006 my $table = $source->from;
1007 @colvalues{@$cols} = (0..$#$cols);
1008 my ($sql, @bind) = $self->sql_maker->insert($table, \%colvalues);
1010 $self->_query_start( $sql, @bind );
1011 my $sth = $self->sth($sql);
1013 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1015 ## This must be an arrayref, else nothing works!
1016 my $tuple_status = [];
1018 ## Get the bind_attributes, if any exist
1019 my $bind_attributes = $self->source_bind_attributes($source);
1021 ## Bind the values and execute
1022 my $placeholder_index = 1;
1024 foreach my $bound (@bind) {
1026 my $attributes = {};
1027 my ($column_name, $data_index) = @$bound;
1029 if( $bind_attributes ) {
1030 $attributes = $bind_attributes->{$column_name}
1031 if defined $bind_attributes->{$column_name};
1034 my @data = map { $_->[$data_index] } @$data;
1036 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1037 $placeholder_index++;
1039 my $rv = $sth->execute_array({ArrayTupleStatus => $tuple_status});
1040 $self->throw_exception($sth->errstr) if !$rv;
1042 $self->_query_end( $sql, @bind );
1043 return (wantarray ? ($rv, $sth, @bind) : $rv);
1047 my $self = shift @_;
1048 my $source = shift @_;
1049 my $bind_attributes = $self->source_bind_attributes($source);
1051 return $self->_execute('update' => [], $source, $bind_attributes, @_);
1056 my $self = shift @_;
1057 my $source = shift @_;
1059 my $bind_attrs = {}; ## If ever it's needed...
1061 return $self->_execute('delete' => [], $source, $bind_attrs, @_);
1064 # We were sent here because the $rs contains a complex search
1065 # which will require a subquery to select the correct rows
1066 # (i.e. joined or limited resultsets)
1068 # Genarating a single PK column subquery is trivial and supported
1069 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1070 # Look at multipk_update_delete()
1071 sub subq_update_delete {
1073 my ($rs, $op, $values) = @_;
1075 if ($rs->result_source->primary_columns == 1) {
1076 return $self->_onepk_update_delete (@_);
1079 return $self->_multipk_update_delete (@_);
1083 # Generally a single PK resultset operation is trivially expressed
1084 # with PK IN (subquery). However some databases (mysql) do not support
1085 # modification of a table mentioned in the subselect. This method
1086 # should be overriden in the appropriate storage class to be smarter
1087 # in such situations
1088 sub _onepk_update_delete {
1091 my ($rs, $op, $values) = @_;
1093 my $rsrc = $rs->result_source;
1094 my $attrs = $rs->_resolved_attrs;
1095 my @pcols = $rsrc->primary_columns;
1097 $self->throw_exception ('_onepk_update_delete can not be called on resultsets selecting multiple columns')
1098 if (ref $attrs->{select} eq 'ARRAY' and @{$attrs->{select}} > 1);
1102 $op eq 'update' ? $values : (),
1103 { $pcols[0] => { -in => $rs->as_query } },
1107 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1108 # resultset update/delete involving subqueries. So resort to simple
1109 # (and inefficient) delete_all style per-row opearations, while allowing
1110 # specific storages to override this with a faster implementation.
1112 # We do not use $row->$op style queries, because resultset update/delete
1113 # is not expected to cascade (this is what delete_all/update_all is for).
1115 # There should be no race conditions as the entire operation is rolled
1117 sub _multipk_update_delete {
1119 my ($rs, $op, $values) = @_;
1121 my $rsrc = $rs->result_source;
1122 my @pcols = $rsrc->primary_columns;
1123 my $attrs = $rs->_resolved_attrs;
1125 $self->throw_exception ('Number of columns selected by supplied resultset does not match number of primary keys')
1126 if ( ref $attrs->{select} ne 'ARRAY' or @{$attrs->{select}} != @pcols );
1128 my $guard = $self->txn_scope_guard;
1130 my $subrs_cur = $rs->cursor;
1131 while (my @pks = $subrs_cur->next) {
1134 for my $i (0.. $#pcols) {
1135 $cond->{$pcols[$i]} = $pks[$i];
1140 $op eq 'update' ? $values : (),
1153 my $sql_maker = $self->sql_maker;
1154 local $sql_maker->{for};
1155 return $self->_execute($self->_select_args(@_));
1159 my ($self, $ident, $select, $condition, $attrs) = @_;
1160 my $order = $attrs->{order_by};
1162 my $for = delete $attrs->{for};
1163 my $sql_maker = $self->sql_maker;
1164 $sql_maker->{for} = $for;
1166 if (exists $attrs->{group_by} || $attrs->{having}) {
1168 group_by => $attrs->{group_by},
1169 having => $attrs->{having},
1170 ($order ? (order_by => $order) : ())
1173 my $bind_attrs = {}; ## Future support
1174 my @args = ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $condition, $order);
1175 if ($attrs->{software_limit} ||
1176 $self->sql_maker->_default_limit_syntax eq "GenericSubQ") {
1177 $attrs->{software_limit} = 1;
1179 $self->throw_exception("rows attribute must be positive if present")
1180 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1182 # MySQL actually recommends this approach. I cringe.
1183 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1184 push @args, $attrs->{rows}, $attrs->{offset};
1189 sub source_bind_attributes {
1190 my ($self, $source) = @_;
1192 my $bind_attributes;
1193 foreach my $column ($source->columns) {
1195 my $data_type = $source->column_info($column)->{data_type} || '';
1196 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
1200 return $bind_attributes;
1207 =item Arguments: $ident, $select, $condition, $attrs
1211 Handle a SQL select statement.
1217 my ($ident, $select, $condition, $attrs) = @_;
1218 return $self->cursor_class->new($self, \@_, $attrs);
1223 my ($rv, $sth, @bind) = $self->_select(@_);
1224 my @row = $sth->fetchrow_array;
1225 my @nextrow = $sth->fetchrow_array if @row;
1226 if(@row && @nextrow) {
1227 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
1229 # Need to call finish() to work round broken DBDs
1238 =item Arguments: $sql
1242 Returns a L<DBI> sth (statement handle) for the supplied SQL.
1247 my ($self, $dbh, $sql) = @_;
1249 # 3 is the if_active parameter which avoids active sth re-use
1250 my $sth = $self->disable_sth_caching
1251 ? $dbh->prepare($sql)
1252 : $dbh->prepare_cached($sql, {}, 3);
1254 # XXX You would think RaiseError would make this impossible,
1255 # but apparently that's not true :(
1256 $self->throw_exception($dbh->errstr) if !$sth;
1262 my ($self, $sql) = @_;
1263 $self->dbh_do('_dbh_sth', $sql);
1266 sub _dbh_columns_info_for {
1267 my ($self, $dbh, $table) = @_;
1269 if ($dbh->can('column_info')) {
1272 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
1273 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
1275 while ( my $info = $sth->fetchrow_hashref() ){
1277 $column_info{data_type} = $info->{TYPE_NAME};
1278 $column_info{size} = $info->{COLUMN_SIZE};
1279 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
1280 $column_info{default_value} = $info->{COLUMN_DEF};
1281 my $col_name = $info->{COLUMN_NAME};
1282 $col_name =~ s/^\"(.*)\"$/$1/;
1284 $result{$col_name} = \%column_info;
1287 return \%result if !$@ && scalar keys %result;
1291 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
1293 my @columns = @{$sth->{NAME_lc}};
1294 for my $i ( 0 .. $#columns ){
1296 $column_info{data_type} = $sth->{TYPE}->[$i];
1297 $column_info{size} = $sth->{PRECISION}->[$i];
1298 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
1300 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
1301 $column_info{data_type} = $1;
1302 $column_info{size} = $2;
1305 $result{$columns[$i]} = \%column_info;
1309 foreach my $col (keys %result) {
1310 my $colinfo = $result{$col};
1311 my $type_num = $colinfo->{data_type};
1313 if(defined $type_num && $dbh->can('type_info')) {
1314 my $type_info = $dbh->type_info($type_num);
1315 $type_name = $type_info->{TYPE_NAME} if $type_info;
1316 $colinfo->{data_type} = $type_name if $type_name;
1323 sub columns_info_for {
1324 my ($self, $table) = @_;
1325 $self->dbh_do('_dbh_columns_info_for', $table);
1328 =head2 last_insert_id
1330 Return the row id of the last insert.
1334 sub _dbh_last_insert_id {
1335 # All Storage's need to register their own _dbh_last_insert_id
1336 # the old SQLite-based method was highly inappropriate
1339 my $class = ref $self;
1340 $self->throw_exception (<<EOE);
1342 No _dbh_last_insert_id() method found in $class.
1343 Since the method of obtaining the autoincrement id of the last insert
1344 operation varies greatly between different databases, this method must be
1345 individually implemented for every storage class.
1349 sub last_insert_id {
1351 $self->dbh_do('_dbh_last_insert_id', @_);
1356 Returns the database driver name.
1360 sub sqlt_type { shift->dbh->{Driver}->{Name} }
1362 =head2 bind_attribute_by_data_type
1364 Given a datatype from column info, returns a database specific bind
1365 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
1366 let the database planner just handle it.
1368 Generally only needed for special case column types, like bytea in postgres.
1372 sub bind_attribute_by_data_type {
1376 =head2 create_ddl_dir
1380 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
1384 Creates a SQL file based on the Schema, for each of the specified
1385 database types, in the given directory.
1387 By default, C<\%sqlt_args> will have
1389 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
1391 merged with the hash passed in. To disable any of those features, pass in a
1392 hashref like the following
1394 { ignore_constraint_names => 0, # ... other options }
1398 sub create_ddl_dir {
1399 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
1401 if(!$dir || !-d $dir) {
1402 carp "No directory given, using ./\n";
1405 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
1406 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
1408 my $schema_version = $schema->schema_version || '1.x';
1409 $version ||= $schema_version;
1412 add_drop_table => 1,
1413 ignore_constraint_names => 1,
1414 ignore_index_names => 1,
1418 $self->throw_exception(q{Can't create a ddl file without SQL::Translator 0.09003: '}
1419 . $self->_check_sqlt_message . q{'})
1420 if !$self->_check_sqlt_version;
1422 my $sqlt = SQL::Translator->new( $sqltargs );
1424 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
1425 my $sqlt_schema = $sqlt->translate({ data => $schema })
1426 or $self->throw_exception ($sqlt->error);
1428 foreach my $db (@$databases) {
1430 $sqlt->{schema} = $sqlt_schema;
1431 $sqlt->producer($db);
1434 my $filename = $schema->ddl_filename($db, $version, $dir);
1435 if (-e $filename && ($version eq $schema_version )) {
1436 # if we are dumping the current version, overwrite the DDL
1437 carp "Overwriting existing DDL file - $filename";
1441 my $output = $sqlt->translate;
1443 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
1446 if(!open($file, ">$filename")) {
1447 $self->throw_exception("Can't open $filename for writing ($!)");
1450 print $file $output;
1453 next unless ($preversion);
1455 require SQL::Translator::Diff;
1457 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
1458 if(!-e $prefilename) {
1459 carp("No previous schema file found ($prefilename)");
1463 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
1465 carp("Overwriting existing diff file - $difffile");
1471 my $t = SQL::Translator->new($sqltargs);
1476 or $self->throw_exception ($t->error);
1478 my $out = $t->translate( $prefilename )
1479 or $self->throw_exception ($t->error);
1481 $source_schema = $t->schema;
1483 $source_schema->name( $prefilename )
1484 unless ( $source_schema->name );
1487 # The "new" style of producers have sane normalization and can support
1488 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
1489 # And we have to diff parsed SQL against parsed SQL.
1490 my $dest_schema = $sqlt_schema;
1492 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
1493 my $t = SQL::Translator->new($sqltargs);
1498 or $self->throw_exception ($t->error);
1500 my $out = $t->translate( $filename )
1501 or $self->throw_exception ($t->error);
1503 $dest_schema = $t->schema;
1505 $dest_schema->name( $filename )
1506 unless $dest_schema->name;
1509 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
1513 if(!open $file, ">$difffile") {
1514 $self->throw_exception("Can't write to $difffile ($!)");
1522 =head2 deployment_statements
1526 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
1530 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
1531 The database driver name is given by C<$type>, though the value from
1532 L</sqlt_type> is used if it is not specified.
1534 C<$directory> is used to return statements from files in a previously created
1535 L</create_ddl_dir> directory and is optional. The filenames are constructed
1536 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
1538 If no C<$directory> is specified then the statements are constructed on the
1539 fly using L<SQL::Translator> and C<$version> is ignored.
1541 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
1545 sub deployment_statements {
1546 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
1547 # Need to be connected to get the correct sqlt_type
1548 $self->ensure_connected() unless $type;
1549 $type ||= $self->sqlt_type;
1550 $version ||= $schema->schema_version || '1.x';
1552 my $filename = $schema->ddl_filename($type, $version, $dir);
1556 open($file, "<$filename")
1557 or $self->throw_exception("Can't open $filename ($!)");
1560 return join('', @rows);
1563 $self->throw_exception(q{Can't deploy without SQL::Translator 0.09003: '}
1564 . $self->_check_sqlt_message . q{'})
1565 if !$self->_check_sqlt_version;
1567 require SQL::Translator::Parser::DBIx::Class;
1568 eval qq{use SQL::Translator::Producer::${type}};
1569 $self->throw_exception($@) if $@;
1571 # sources needs to be a parser arg, but for simplicty allow at top level
1573 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
1574 if exists $sqltargs->{sources};
1576 my $tr = SQL::Translator->new(%$sqltargs);
1577 SQL::Translator::Parser::DBIx::Class::parse( $tr, $schema );
1578 return "SQL::Translator::Producer::${type}"->can('produce')->($tr);
1582 my ($self, $schema, $type, $sqltargs, $dir) = @_;
1585 return if($line =~ /^--/);
1587 # next if($line =~ /^DROP/m);
1588 return if($line =~ /^BEGIN TRANSACTION/m);
1589 return if($line =~ /^COMMIT/m);
1590 return if $line =~ /^\s+$/; # skip whitespace only
1591 $self->_query_start($line);
1593 $self->dbh->do($line); # shouldn't be using ->dbh ?
1596 carp qq{$@ (running "${line}")};
1598 $self->_query_end($line);
1600 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { no_comments => 1, %{ $sqltargs || {} } } );
1601 if (@statements > 1) {
1602 foreach my $statement (@statements) {
1603 $deploy->( $statement );
1606 elsif (@statements == 1) {
1607 foreach my $line ( split(";\n", $statements[0])) {
1613 =head2 datetime_parser
1615 Returns the datetime parser class
1619 sub datetime_parser {
1621 return $self->{datetime_parser} ||= do {
1622 $self->ensure_connected;
1623 $self->build_datetime_parser(@_);
1627 =head2 datetime_parser_type
1629 Defines (returns) the datetime parser class - currently hardwired to
1630 L<DateTime::Format::MySQL>
1634 sub datetime_parser_type { "DateTime::Format::MySQL"; }
1636 =head2 build_datetime_parser
1638 See L</datetime_parser>
1642 sub build_datetime_parser {
1644 my $type = $self->datetime_parser_type(@_);
1646 $self->throw_exception("Couldn't load ${type}: $@") if $@;
1651 my $_check_sqlt_version; # private
1652 my $_check_sqlt_message; # private
1653 sub _check_sqlt_version {
1654 return $_check_sqlt_version if defined $_check_sqlt_version;
1655 eval 'use SQL::Translator "0.09003"';
1656 $_check_sqlt_message = $@ || '';
1657 $_check_sqlt_version = !$@;
1660 sub _check_sqlt_message {
1661 _check_sqlt_version if !defined $_check_sqlt_message;
1662 $_check_sqlt_message;
1666 =head2 is_replicating
1668 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
1669 replicate from a master database. Default is undef, which is the result
1670 returned by databases that don't support replication.
1674 sub is_replicating {
1679 =head2 lag_behind_master
1681 Returns a number that represents a certain amount of lag behind a master db
1682 when a given storage is replicating. The number is database dependent, but
1683 starts at zero and increases with the amount of lag. Default in undef
1687 sub lag_behind_master {
1693 return if !$self->_dbh;
1702 =head2 DBIx::Class and AutoCommit
1704 DBIx::Class can do some wonderful magic with handling exceptions,
1705 disconnections, and transactions when you use C<< AutoCommit => 1 >>
1706 combined with C<txn_do> for transaction support.
1708 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
1709 in an assumed transaction between commits, and you're telling us you'd
1710 like to manage that manually. A lot of the magic protections offered by
1711 this module will go away. We can't protect you from exceptions due to database
1712 disconnects because we don't know anything about how to restart your
1713 transactions. You're on your own for handling all sorts of exceptional
1714 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
1721 Matt S. Trout <mst@shadowcatsystems.co.uk>
1723 Andy Grundman <andy@hybridized.org>
1727 You may distribute this code under the same terms as Perl itself.