1 package DBIx::Class::Storage::DBI;
2 # -*- mode: cperl; cperl-indent-level: 2 -*-
7 use base qw/DBIx::Class::Storage::DBIHacks DBIx::Class::Storage/;
10 use Carp::Clan qw/^DBIx::Class|^Try::Tiny/;
12 use DBIx::Class::Storage::DBI::Cursor;
13 use DBIx::Class::Storage::Statistics;
14 use Scalar::Util qw/refaddr weaken reftype blessed/;
15 use Data::Dumper::Concise 'Dumper';
16 use Sub::Name 'subname';
18 use File::Path 'make_path';
22 # default cursor class, overridable in connect_info attributes
23 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
25 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class sql_limit_dialect/);
26 __PACKAGE__->sql_maker_class('DBIx::Class::SQLMaker');
28 __PACKAGE__->mk_group_accessors('simple' => qw/
29 _connect_info _dbi_connect_info _dbic_connect_attributes _driver_determined
30 _dbh _dbh_details _conn_pid _conn_tid _sql_maker _sql_maker_opts
31 transaction_depth _dbh_autocommit savepoints
34 # the values for these accessors are picked out (and deleted) from
35 # the attribute hashref passed to connect_info
36 my @storage_options = qw/
37 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
38 disable_sth_caching unsafe auto_savepoint
40 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
43 # capability definitions, using a 2-tiered accessor system
46 # A driver/user may define _use_X, which blindly without any checks says:
47 # "(do not) use this capability", (use_dbms_capability is an "inherited"
50 # If _use_X is undef, _supports_X is then queried. This is a "simple" style
51 # accessor, which in turn calls _determine_supports_X, and stores the return
52 # in a special slot on the storage object, which is wiped every time a $dbh
53 # reconnection takes place (it is not guaranteed that upon reconnection we
54 # will get the same rdbms version). _determine_supports_X does not need to
55 # exist on a driver, as we ->can for it before calling.
57 my @capabilities = (qw/insert_returning placeholders typeless_placeholders/);
58 __PACKAGE__->mk_group_accessors( dbms_capability => map { "_supports_$_" } @capabilities );
59 __PACKAGE__->mk_group_accessors( use_dbms_capability => map { "_use_$_" } @capabilities );
62 # Each of these methods need _determine_driver called before itself
63 # in order to function reliably. This is a purely DRY optimization
65 # get_(use)_dbms_capability need to be called on the correct Storage
66 # class, as _use_X may be hardcoded class-wide, and _supports_X calls
67 # _determine_supports_X which obv. needs a correct driver as well
68 my @rdbms_specific_methods = qw/
82 get_use_dbms_capability
89 for my $meth (@rdbms_specific_methods) {
91 my $orig = __PACKAGE__->can ($meth)
92 or die "$meth is not a ::Storage::DBI method!";
95 no warnings qw/redefine/;
96 *{__PACKAGE__ ."::$meth"} = subname $meth => sub {
97 if (not $_[0]->_driver_determined and not $_[0]->{_in_determine_driver}) {
98 $_[0]->_determine_driver;
100 # This for some reason crashes and burns on perl 5.8.1
101 # IFF the method ends up throwing an exception
102 #goto $_[0]->can ($meth);
104 my $cref = $_[0]->can ($meth);
114 DBIx::Class::Storage::DBI - DBI storage handler
118 my $schema = MySchema->connect('dbi:SQLite:my.db');
120 $schema->storage->debug(1);
122 my @stuff = $schema->storage->dbh_do(
124 my ($storage, $dbh, @args) = @_;
125 $dbh->do("DROP TABLE authors");
130 $schema->resultset('Book')->search({
131 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
136 This class represents the connection to an RDBMS via L<DBI>. See
137 L<DBIx::Class::Storage> for general information. This pod only
138 documents DBI-specific methods and behaviors.
145 my $new = shift->next::method(@_);
147 $new->transaction_depth(0);
148 $new->_sql_maker_opts({});
149 $new->_dbh_details({});
150 $new->{savepoints} = [];
151 $new->{_in_dbh_do} = 0;
152 $new->{_dbh_gen} = 0;
154 # read below to see what this does
155 $new->_arm_global_destructor;
160 # This is hack to work around perl shooting stuff in random
161 # order on exit(). If we do not walk the remaining storage
162 # objects in an END block, there is a *small but real* chance
163 # of a fork()ed child to kill the parent's shared DBI handle,
164 # *before perl reaches the DESTROY in this package*
165 # Yes, it is ugly and effective.
167 my %seek_and_destroy;
169 sub _arm_global_destructor {
171 my $key = Scalar::Util::refaddr ($self);
172 $seek_and_destroy{$key} = $self;
173 Scalar::Util::weaken ($seek_and_destroy{$key});
177 local $?; # just in case the DBI destructor changes it somehow
179 # destroy just the object if not native to this process/thread
180 $_->_preserve_foreign_dbh for (grep
182 values %seek_and_destroy
190 # destroy just the object if not native to this process/thread
191 $self->_preserve_foreign_dbh;
193 # some databases need this to stop spewing warnings
194 if (my $dbh = $self->_dbh) {
196 %{ $dbh->{CachedKids} } = ();
204 sub _preserve_foreign_dbh {
207 return unless $self->_dbh;
211 return unless $self->_dbh;
217 # handle pid changes correctly - do not destroy parent's connection
221 return if ( defined $self->_conn_pid and $self->_conn_pid == $$ );
223 $self->_dbh->{InactiveDestroy} = 1;
230 # very similar to above, but seems to FAIL if I set InactiveDestroy
234 if ( ! defined $self->_conn_tid ) {
237 elsif ( $self->_conn_tid == threads->tid ) {
238 return; # same thread
241 #$self->_dbh->{InactiveDestroy} = 1; # why does t/51threads.t fail...?
251 This method is normally called by L<DBIx::Class::Schema/connection>, which
252 encapsulates its argument list in an arrayref before passing them here.
254 The argument list may contain:
260 The same 4-element argument set one would normally pass to
261 L<DBI/connect>, optionally followed by
262 L<extra attributes|/DBIx::Class specific connection attributes>
263 recognized by DBIx::Class:
265 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
269 A single code reference which returns a connected
270 L<DBI database handle|DBI/connect> optionally followed by
271 L<extra attributes|/DBIx::Class specific connection attributes> recognized
274 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
278 A single hashref with all the attributes and the dsn/user/password
281 $connect_info_args = [{
289 $connect_info_args = [{
290 dbh_maker => sub { DBI->connect (...) },
295 This is particularly useful for L<Catalyst> based applications, allowing the
296 following config (L<Config::General> style):
301 dsn dbi:mysql:database=test
308 The C<dsn>/C<user>/C<password> combination can be substituted by the
309 C<dbh_maker> key whose value is a coderef that returns a connected
310 L<DBI database handle|DBI/connect>
314 Please note that the L<DBI> docs recommend that you always explicitly
315 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
316 recommends that it be set to I<1>, and that you perform transactions
317 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
318 to I<1> if you do not do explicitly set it to zero. This is the default
319 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
321 =head3 DBIx::Class specific connection attributes
323 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
324 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
325 the following connection options. These options can be mixed in with your other
326 L<DBI> connection attributes, or placed in a separate hashref
327 (C<\%extra_attributes>) as shown above.
329 Every time C<connect_info> is invoked, any previous settings for
330 these options will be cleared before setting the new ones, regardless of
331 whether any options are specified in the new C<connect_info>.
338 Specifies things to do immediately after connecting or re-connecting to
339 the database. Its value may contain:
345 This contains one SQL statement to execute.
347 =item an array reference
349 This contains SQL statements to execute in order. Each element contains
350 a string or a code reference that returns a string.
352 =item a code reference
354 This contains some code to execute. Unlike code references within an
355 array reference, its return value is ignored.
359 =item on_disconnect_do
361 Takes arguments in the same form as L</on_connect_do> and executes them
362 immediately before disconnecting from the database.
364 Note, this only runs if you explicitly call L</disconnect> on the
367 =item on_connect_call
369 A more generalized form of L</on_connect_do> that calls the specified
370 C<connect_call_METHOD> methods in your storage driver.
372 on_connect_do => 'select 1'
376 on_connect_call => [ [ do_sql => 'select 1' ] ]
378 Its values may contain:
384 Will call the C<connect_call_METHOD> method.
386 =item a code reference
388 Will execute C<< $code->($storage) >>
390 =item an array reference
392 Each value can be a method name or code reference.
394 =item an array of arrays
396 For each array, the first item is taken to be the C<connect_call_> method name
397 or code reference, and the rest are parameters to it.
401 Some predefined storage methods you may use:
407 Executes a SQL string or a code reference that returns a SQL string. This is
408 what L</on_connect_do> and L</on_disconnect_do> use.
416 Will execute the scalar as SQL.
420 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
421 attributes hashref and bind values.
423 =item a code reference
425 Will execute C<< $code->($storage) >> and execute the return array refs as
432 Execute any statements necessary to initialize the database session to return
433 and accept datetime/timestamp values used with
434 L<DBIx::Class::InflateColumn::DateTime>.
436 Only necessary for some databases, see your specific storage driver for
437 implementation details.
441 =item on_disconnect_call
443 Takes arguments in the same form as L</on_connect_call> and executes them
444 immediately before disconnecting from the database.
446 Calls the C<disconnect_call_METHOD> methods as opposed to the
447 C<connect_call_METHOD> methods called by L</on_connect_call>.
449 Note, this only runs if you explicitly call L</disconnect> on the
452 =item disable_sth_caching
454 If set to a true value, this option will disable the caching of
455 statement handles via L<DBI/prepare_cached>.
459 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
460 default L</sql_limit_dialect> setting of the storage (if any). For a list
461 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
465 Specifies what characters to use to quote table and column names. If
466 you use this you will want to specify L</name_sep> as well.
468 C<quote_char> expects either a single character, in which case is it
469 is placed on either side of the table/column name, or an arrayref of length
470 2 in which case the table/column name is placed between the elements.
472 For example under MySQL you should use C<< quote_char => '`' >>, and for
473 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
477 This only needs to be used in conjunction with C<quote_char>, and is used to
478 specify the character that separates elements (schemas, tables, columns) from
479 each other. In most cases this is simply a C<.>.
481 The consequences of not supplying this value is that L<SQL::Abstract>
482 will assume DBIx::Class' uses of aliases to be complete column
483 names. The output will look like I<"me.name"> when it should actually
488 This Storage driver normally installs its own C<HandleError>, sets
489 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
490 all database handles, including those supplied by a coderef. It does this
491 so that it can have consistent and useful error behavior.
493 If you set this option to a true value, Storage will not do its usual
494 modifications to the database handle's attributes, and instead relies on
495 the settings in your connect_info DBI options (or the values you set in
496 your connection coderef, in the case that you are connecting via coderef).
498 Note that your custom settings can cause Storage to malfunction,
499 especially if you set a C<HandleError> handler that suppresses exceptions
500 and/or disable C<RaiseError>.
504 If this option is true, L<DBIx::Class> will use savepoints when nesting
505 transactions, making it possible to recover from failure in the inner
506 transaction without having to abort all outer transactions.
510 Use this argument to supply a cursor class other than the default
511 L<DBIx::Class::Storage::DBI::Cursor>.
515 Some real-life examples of arguments to L</connect_info> and
516 L<DBIx::Class::Schema/connect>
518 # Simple SQLite connection
519 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
522 ->connect_info([ sub { DBI->connect(...) } ]);
524 # Connect via subref in hashref
526 dbh_maker => sub { DBI->connect(...) },
527 on_connect_do => 'alter session ...',
530 # A bit more complicated
537 { quote_char => q{"}, name_sep => q{.} },
541 # Equivalent to the previous example
547 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
551 # Same, but with hashref as argument
552 # See parse_connect_info for explanation
555 dsn => 'dbi:Pg:dbname=foo',
557 password => 'my_pg_password',
564 # Subref + DBIx::Class-specific connection options
567 sub { DBI->connect(...) },
571 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
572 disable_sth_caching => 1,
582 my ($self, $info) = @_;
584 return $self->_connect_info if !$info;
586 $self->_connect_info($info); # copy for _connect_info
588 $info = $self->_normalize_connect_info($info)
589 if ref $info eq 'ARRAY';
591 for my $storage_opt (keys %{ $info->{storage_options} }) {
592 my $value = $info->{storage_options}{$storage_opt};
594 $self->$storage_opt($value);
597 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
598 # the new set of options
599 $self->_sql_maker(undef);
600 $self->_sql_maker_opts({});
602 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
603 my $value = $info->{sql_maker_options}{$sql_maker_opt};
605 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
609 %{ $self->_default_dbi_connect_attributes || {} },
610 %{ $info->{attributes} || {} },
613 my @args = @{ $info->{arguments} };
615 $self->_dbi_connect_info([@args,
616 %attrs && !(ref $args[0] eq 'CODE') ? \%attrs : ()]);
619 # save attributes them in a separate accessor so they are always
620 # introspectable, even in case of a CODE $dbhmaker
621 $self->_dbic_connect_attributes (\%attrs);
623 return $self->_connect_info;
626 sub _normalize_connect_info {
627 my ($self, $info_arg) = @_;
630 my @args = @$info_arg; # take a shallow copy for further mutilation
632 # combine/pre-parse arguments depending on invocation style
635 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
636 %attrs = %{ $args[1] || {} };
639 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
640 %attrs = %{$args[0]};
642 if (my $code = delete $attrs{dbh_maker}) {
645 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
648 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
649 . "to the result of 'dbh_maker'",
651 join (', ', map { "'$_'" } (@ignored) ),
656 @args = delete @attrs{qw/dsn user password/};
659 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
661 % { $args[3] || {} },
662 % { $args[4] || {} },
664 @args = @args[0,1,2];
667 $info{arguments} = \@args;
669 my @storage_opts = grep exists $attrs{$_},
670 @storage_options, 'cursor_class';
672 @{ $info{storage_options} }{@storage_opts} =
673 delete @attrs{@storage_opts} if @storage_opts;
675 my @sql_maker_opts = grep exists $attrs{$_},
676 qw/limit_dialect quote_char name_sep/;
678 @{ $info{sql_maker_options} }{@sql_maker_opts} =
679 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
681 $info{attributes} = \%attrs if %attrs;
686 sub _default_dbi_connect_attributes {
696 This method is deprecated in favour of setting via L</connect_info>.
700 =head2 on_disconnect_do
702 This method is deprecated in favour of setting via L</connect_info>.
706 sub _parse_connect_do {
707 my ($self, $type) = @_;
709 my $val = $self->$type;
710 return () if not defined $val;
715 push @res, [ 'do_sql', $val ];
716 } elsif (ref($val) eq 'CODE') {
718 } elsif (ref($val) eq 'ARRAY') {
719 push @res, map { [ 'do_sql', $_ ] } @$val;
721 $self->throw_exception("Invalid type for $type: ".ref($val));
729 Arguments: ($subref | $method_name), @extra_coderef_args?
731 Execute the given $subref or $method_name using the new exception-based
732 connection management.
734 The first two arguments will be the storage object that C<dbh_do> was called
735 on and a database handle to use. Any additional arguments will be passed
736 verbatim to the called subref as arguments 2 and onwards.
738 Using this (instead of $self->_dbh or $self->dbh) ensures correct
739 exception handling and reconnection (or failover in future subclasses).
741 Your subref should have no side-effects outside of the database, as
742 there is the potential for your subref to be partially double-executed
743 if the database connection was stale/dysfunctional.
747 my @stuff = $schema->storage->dbh_do(
749 my ($storage, $dbh, @cols) = @_;
750 my $cols = join(q{, }, @cols);
751 $dbh->selectrow_array("SELECT $cols FROM foo");
762 my $dbh = $self->_get_dbh;
764 return $self->$code($dbh, @_)
765 if ( $self->{_in_dbh_do} || $self->{transaction_depth} );
767 local $self->{_in_dbh_do} = 1;
769 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
772 $self->$code ($dbh, @$args);
774 $self->throw_exception($_) if $self->connected;
776 # We were not connected - reconnect and retry, but let any
777 # exception fall right through this time
778 carp "Retrying $code after catching disconnected exception: $_"
779 if $ENV{DBIC_DBIRETRY_DEBUG};
781 $self->_populate_dbh;
782 $self->$code($self->_dbh, @$args);
786 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
787 # It also informs dbh_do to bypass itself while under the direction of txn_do,
788 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
793 ref $coderef eq 'CODE' or $self->throw_exception
794 ('$coderef must be a CODE reference');
796 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
798 local $self->{_in_dbh_do} = 1;
801 my $want_array = wantarray;
807 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
813 @result = $coderef->(@$args);
815 elsif(defined $want_array) {
816 $result[0] = $coderef->(@$args);
826 if(! defined $exception) { return $want_array ? @result : $result[0] }
828 if($tried++ || $self->connected) {
829 my $rollback_exception;
830 try { $self->txn_rollback } catch { $rollback_exception = shift };
831 if(defined $rollback_exception) {
832 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
833 $self->throw_exception($exception) # propagate nested rollback
834 if $rollback_exception =~ /$exception_class/;
836 $self->throw_exception(
837 "Transaction aborted: ${exception}. "
838 . "Rollback failed: ${rollback_exception}"
841 $self->throw_exception($exception)
844 # We were not connected, and was first try - reconnect and retry
846 carp "Retrying $coderef after catching disconnected exception: $exception"
847 if $ENV{DBIC_TXNRETRY_DEBUG};
848 $self->_populate_dbh;
854 Our C<disconnect> method also performs a rollback first if the
855 database is not in C<AutoCommit> mode.
865 push @actions, ( $self->on_disconnect_call || () );
866 push @actions, $self->_parse_connect_do ('on_disconnect_do');
868 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
870 $self->_dbh_rollback unless $self->_dbh_autocommit;
872 %{ $self->_dbh->{CachedKids} } = ();
873 $self->_dbh->disconnect;
879 =head2 with_deferred_fk_checks
883 =item Arguments: C<$coderef>
885 =item Return Value: The return value of $coderef
889 Storage specific method to run the code ref with FK checks deferred or
890 in MySQL's case disabled entirely.
894 # Storage subclasses should override this
895 sub with_deferred_fk_checks {
896 my ($self, $sub) = @_;
904 =item Arguments: none
906 =item Return Value: 1|0
910 Verifies that the current database handle is active and ready to execute
911 an SQL statement (e.g. the connection did not get stale, server is still
912 answering, etc.) This method is used internally by L</dbh>.
918 return 0 unless $self->_seems_connected;
921 local $self->_dbh->{RaiseError} = 1;
926 sub _seems_connected {
929 $self->_preserve_foreign_dbh;
931 my $dbh = $self->_dbh
934 return $dbh->FETCH('Active');
940 my $dbh = $self->_dbh or return 0;
945 sub ensure_connected {
948 unless ($self->connected) {
949 $self->_populate_dbh;
955 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
956 is guaranteed to be healthy by implicitly calling L</connected>, and if
957 necessary performing a reconnection before returning. Keep in mind that this
958 is very B<expensive> on some database engines. Consider using L</dbh_do>
966 if (not $self->_dbh) {
967 $self->_populate_dbh;
969 $self->ensure_connected;
974 # this is the internal "get dbh or connect (don't check)" method
977 $self->_preserve_foreign_dbh;
978 $self->_populate_dbh unless $self->_dbh;
984 unless ($self->_sql_maker) {
985 my $sql_maker_class = $self->sql_maker_class;
986 $self->ensure_class_loaded ($sql_maker_class);
988 my %opts = %{$self->_sql_maker_opts||{}};
992 $self->sql_limit_dialect
995 my $s_class = (ref $self) || $self;
997 "Your storage class ($s_class) does not set sql_limit_dialect and you "
998 . 'have not supplied an explicit limit_dialect in your connection_info. '
999 . 'DBIC will attempt to use the GenericSubQ dialect, which works on most '
1000 . 'databases but can be (and often is) painfully slow.'
1007 $self->_sql_maker($sql_maker_class->new(
1008 bindtype=>'columns',
1009 array_datatypes => 1,
1010 limit_dialect => $dialect,
1014 return $self->_sql_maker;
1017 # nothing to do by default
1024 my @info = @{$self->_dbi_connect_info || []};
1025 $self->_dbh(undef); # in case ->connected failed we might get sent here
1026 $self->_dbh_details({}); # reset everything we know
1028 $self->_dbh($self->_connect(@info));
1030 $self->_conn_pid($$);
1031 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
1033 $self->_determine_driver;
1035 # Always set the transaction depth on connect, since
1036 # there is no transaction in progress by definition
1037 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1039 $self->_run_connection_actions unless $self->{_in_determine_driver};
1042 sub _run_connection_actions {
1046 push @actions, ( $self->on_connect_call || () );
1047 push @actions, $self->_parse_connect_do ('on_connect_do');
1049 $self->_do_connection_actions(connect_call_ => $_) for @actions;
1054 sub set_use_dbms_capability {
1055 $_[0]->set_inherited ($_[1], $_[2]);
1058 sub get_use_dbms_capability {
1059 my ($self, $capname) = @_;
1061 my $use = $self->get_inherited ($capname);
1064 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1068 sub set_dbms_capability {
1069 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1072 sub get_dbms_capability {
1073 my ($self, $capname) = @_;
1075 my $cap = $self->_dbh_details->{capability}{$capname};
1077 unless (defined $cap) {
1078 if (my $meth = $self->can ("_determine$capname")) {
1079 $cap = $self->$meth ? 1 : 0;
1085 $self->set_dbms_capability ($capname, $cap);
1095 unless ($info = $self->_dbh_details->{info}) {
1099 my $server_version = try { $self->_get_server_version };
1101 if (defined $server_version) {
1102 $info->{dbms_version} = $server_version;
1104 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1105 my @verparts = split (/\./, $numeric_version);
1111 # consider only up to 3 version parts, iff not more than 3 digits
1113 while (@verparts && @use_parts < 3) {
1114 my $p = shift @verparts;
1116 push @use_parts, $p;
1118 push @use_parts, 0 while @use_parts < 3;
1120 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1124 $self->_dbh_details->{info} = $info;
1130 sub _get_server_version {
1131 shift->_get_dbh->get_info(18);
1134 sub _determine_driver {
1137 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1138 my $started_connected = 0;
1139 local $self->{_in_determine_driver} = 1;
1141 if (ref($self) eq __PACKAGE__) {
1143 if ($self->_dbh) { # we are connected
1144 $driver = $self->_dbh->{Driver}{Name};
1145 $started_connected = 1;
1147 # if connect_info is a CODEREF, we have no choice but to connect
1148 if (ref $self->_dbi_connect_info->[0] &&
1149 reftype $self->_dbi_connect_info->[0] eq 'CODE') {
1150 $self->_populate_dbh;
1151 $driver = $self->_dbh->{Driver}{Name};
1154 # try to use dsn to not require being connected, the driver may still
1155 # force a connection in _rebless to determine version
1156 # (dsn may not be supplied at all if all we do is make a mock-schema)
1157 my $dsn = $self->_dbi_connect_info->[0] || $ENV{DBI_DSN} || '';
1158 ($driver) = $dsn =~ /dbi:([^:]+):/i;
1159 $driver ||= $ENV{DBI_DRIVER};
1164 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1165 if ($self->load_optional_class($storage_class)) {
1166 mro::set_mro($storage_class, 'c3');
1167 bless $self, $storage_class;
1173 $self->_driver_determined(1);
1175 $self->_init; # run driver-specific initializations
1177 $self->_run_connection_actions
1178 if !$started_connected && defined $self->_dbh;
1182 sub _do_connection_actions {
1184 my $method_prefix = shift;
1187 if (not ref($call)) {
1188 my $method = $method_prefix . $call;
1190 } elsif (ref($call) eq 'CODE') {
1192 } elsif (ref($call) eq 'ARRAY') {
1193 if (ref($call->[0]) ne 'ARRAY') {
1194 $self->_do_connection_actions($method_prefix, $_) for @$call;
1196 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1199 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1205 sub connect_call_do_sql {
1207 $self->_do_query(@_);
1210 sub disconnect_call_do_sql {
1212 $self->_do_query(@_);
1215 # override in db-specific backend when necessary
1216 sub connect_call_datetime_setup { 1 }
1219 my ($self, $action) = @_;
1221 if (ref $action eq 'CODE') {
1222 $action = $action->($self);
1223 $self->_do_query($_) foreach @$action;
1226 # Most debuggers expect ($sql, @bind), so we need to exclude
1227 # the attribute hash which is the second argument to $dbh->do
1228 # furthermore the bind values are usually to be presented
1229 # as named arrayref pairs, so wrap those here too
1230 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1231 my $sql = shift @do_args;
1232 my $attrs = shift @do_args;
1233 my @bind = map { [ undef, $_ ] } @do_args;
1235 $self->_query_start($sql, @bind);
1236 $self->_get_dbh->do($sql, $attrs, @do_args);
1237 $self->_query_end($sql, @bind);
1244 my ($self, @info) = @_;
1246 $self->throw_exception("You failed to provide any connection info")
1249 my ($old_connect_via, $dbh);
1251 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
1252 $old_connect_via = $DBI::connect_via;
1253 $DBI::connect_via = 'connect';
1257 if(ref $info[0] eq 'CODE') {
1258 $dbh = $info[0]->();
1261 $dbh = DBI->connect(@info);
1268 unless ($self->unsafe) {
1270 # this odd anonymous coderef dereference is in fact really
1271 # necessary to avoid the unwanted effect described in perl5
1274 my $weak_self = $_[0];
1277 $_[1]->{HandleError} = sub {
1279 $weak_self->throw_exception("DBI Exception: $_[0]");
1282 # the handler may be invoked by something totally out of
1284 croak ("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1289 $dbh->{ShowErrorStatement} = 1;
1290 $dbh->{RaiseError} = 1;
1291 $dbh->{PrintError} = 0;
1295 $self->throw_exception("DBI Connection failed: $_")
1298 $DBI::connect_via = $old_connect_via if $old_connect_via;
1301 $self->_dbh_autocommit($dbh->{AutoCommit});
1306 my ($self, $name) = @_;
1308 $name = $self->_svp_generate_name
1309 unless defined $name;
1311 $self->throw_exception ("You can't use savepoints outside a transaction")
1312 if $self->{transaction_depth} == 0;
1314 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1315 unless $self->can('_svp_begin');
1317 push @{ $self->{savepoints} }, $name;
1319 $self->debugobj->svp_begin($name) if $self->debug;
1321 return $self->_svp_begin($name);
1325 my ($self, $name) = @_;
1327 $self->throw_exception ("You can't use savepoints outside a transaction")
1328 if $self->{transaction_depth} == 0;
1330 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1331 unless $self->can('_svp_release');
1333 if (defined $name) {
1334 $self->throw_exception ("Savepoint '$name' does not exist")
1335 unless grep { $_ eq $name } @{ $self->{savepoints} };
1337 # Dig through the stack until we find the one we are releasing. This keeps
1338 # the stack up to date.
1341 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1343 $name = pop @{ $self->{savepoints} };
1346 $self->debugobj->svp_release($name) if $self->debug;
1348 return $self->_svp_release($name);
1352 my ($self, $name) = @_;
1354 $self->throw_exception ("You can't use savepoints outside a transaction")
1355 if $self->{transaction_depth} == 0;
1357 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1358 unless $self->can('_svp_rollback');
1360 if (defined $name) {
1361 # If they passed us a name, verify that it exists in the stack
1362 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1363 $self->throw_exception("Savepoint '$name' does not exist!");
1366 # Dig through the stack until we find the one we are releasing. This keeps
1367 # the stack up to date.
1368 while(my $s = pop(@{ $self->{savepoints} })) {
1369 last if($s eq $name);
1371 # Add the savepoint back to the stack, as a rollback doesn't remove the
1372 # named savepoint, only everything after it.
1373 push(@{ $self->{savepoints} }, $name);
1375 # We'll assume they want to rollback to the last savepoint
1376 $name = $self->{savepoints}->[-1];
1379 $self->debugobj->svp_rollback($name) if $self->debug;
1381 return $self->_svp_rollback($name);
1384 sub _svp_generate_name {
1387 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1393 # this means we have not yet connected and do not know the AC status
1394 # (e.g. coderef $dbh)
1395 $self->ensure_connected if (! defined $self->_dbh_autocommit);
1397 if($self->{transaction_depth} == 0) {
1398 $self->debugobj->txn_begin()
1400 $self->_dbh_begin_work;
1402 elsif ($self->auto_savepoint) {
1405 $self->{transaction_depth}++;
1408 sub _dbh_begin_work {
1411 # if the user is utilizing txn_do - good for him, otherwise we need to
1412 # ensure that the $dbh is healthy on BEGIN.
1413 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1414 # will be replaced by a failure of begin_work itself (which will be
1415 # then retried on reconnect)
1416 if ($self->{_in_dbh_do}) {
1417 $self->_dbh->begin_work;
1419 $self->dbh_do(sub { $_[1]->begin_work });
1425 if ($self->{transaction_depth} == 1) {
1426 $self->debugobj->txn_commit()
1429 $self->{transaction_depth} = 0
1430 if $self->_dbh_autocommit;
1432 elsif($self->{transaction_depth} > 1) {
1433 $self->{transaction_depth}--;
1435 if $self->auto_savepoint;
1441 my $dbh = $self->_dbh
1442 or $self->throw_exception('cannot COMMIT on a disconnected handle');
1448 my $dbh = $self->_dbh;
1450 if ($self->{transaction_depth} == 1) {
1451 $self->debugobj->txn_rollback()
1453 $self->{transaction_depth} = 0
1454 if $self->_dbh_autocommit;
1455 $self->_dbh_rollback;
1457 elsif($self->{transaction_depth} > 1) {
1458 $self->{transaction_depth}--;
1459 if ($self->auto_savepoint) {
1460 $self->svp_rollback;
1465 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1469 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1471 if ($_ !~ /$exception_class/) {
1472 # ensure that a failed rollback resets the transaction depth
1473 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1476 $self->throw_exception($_)
1482 my $dbh = $self->_dbh
1483 or $self->throw_exception('cannot ROLLBACK on a disconnected handle');
1487 # This used to be the top-half of _execute. It was split out to make it
1488 # easier to override in NoBindVars without duping the rest. It takes up
1489 # all of _execute's args, and emits $sql, @bind.
1490 sub _prep_for_execute {
1491 my ($self, $op, $extra_bind, $ident, $args) = @_;
1493 if( blessed $ident && $ident->isa("DBIx::Class::ResultSource") ) {
1494 $ident = $ident->from();
1497 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1500 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1502 return ($sql, \@bind);
1506 sub _fix_bind_params {
1507 my ($self, @bind) = @_;
1509 ### Turn @bind from something like this:
1510 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1512 ### ( "'1'", "'1'", "'3'" )
1515 if ( defined( $_ && $_->[1] ) ) {
1516 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1523 my ( $self, $sql, @bind ) = @_;
1525 if ( $self->debug ) {
1526 @bind = $self->_fix_bind_params(@bind);
1528 $self->debugobj->query_start( $sql, @bind );
1533 my ( $self, $sql, @bind ) = @_;
1535 if ( $self->debug ) {
1536 @bind = $self->_fix_bind_params(@bind);
1537 $self->debugobj->query_end( $sql, @bind );
1542 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1544 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1546 $self->_query_start( $sql, @$bind );
1548 my $sth = $self->sth($sql,$op);
1550 my $placeholder_index = 1;
1552 foreach my $bound (@$bind) {
1553 my $attributes = {};
1554 my($column_name, @data) = @$bound;
1556 if ($bind_attributes) {
1557 $attributes = $bind_attributes->{$column_name}
1558 if defined $bind_attributes->{$column_name};
1561 foreach my $data (@data) {
1562 my $ref = ref $data;
1563 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1565 $sth->bind_param($placeholder_index, $data, $attributes);
1566 $placeholder_index++;
1570 # Can this fail without throwing an exception anyways???
1571 my $rv = $sth->execute();
1572 $self->throw_exception(
1573 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1576 $self->_query_end( $sql, @$bind );
1578 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1583 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1586 sub _prefetch_insert_auto_nextvals {
1587 my ($self, $source, $to_insert) = @_;
1591 foreach my $col ( $source->columns ) {
1592 if ( !defined $to_insert->{$col} ) {
1593 my $col_info = $source->column_info($col);
1595 if ( $col_info->{auto_nextval} ) {
1596 $upd->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1598 $col_info->{sequence} ||=
1599 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1610 my ($source, $to_insert, $opts) = @_;
1612 my $updated_cols = $self->_prefetch_insert_auto_nextvals (@_);
1614 my $bind_attributes = $self->source_bind_attributes($source);
1616 my ($rv, $sth) = $self->_execute('insert' => [], $source, $bind_attributes, $to_insert, $opts);
1618 if ($opts->{returning}) {
1619 my @ret_cols = @{$opts->{returning}};
1621 my @ret_vals = try {
1622 local $SIG{__WARN__} = sub {};
1623 my @r = $sth->fetchrow_array;
1629 @ret{@ret_cols} = @ret_vals if (@ret_vals);
1637 return $updated_cols;
1640 ## Currently it is assumed that all values passed will be "normal", i.e. not
1641 ## scalar refs, or at least, all the same type as the first set, the statement is
1642 ## only prepped once.
1644 my ($self, $source, $cols, $data) = @_;
1647 @colvalues{@$cols} = (0..$#$cols);
1649 for my $i (0..$#$cols) {
1650 my $first_val = $data->[0][$i];
1651 next unless ref $first_val eq 'SCALAR';
1653 $colvalues{ $cols->[$i] } = $first_val;
1656 # check for bad data and stringify stringifiable objects
1657 my $bad_slice = sub {
1658 my ($msg, $col_idx, $slice_idx) = @_;
1659 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
1663 local $Data::Dumper::Maxdepth = 1; # don't dump objects, if any
1665 map { $cols->[$_] => $data->[$slice_idx][$_] } (0 .. $#$cols)
1671 for my $datum_idx (0..$#$data) {
1672 my $datum = $data->[$datum_idx];
1674 for my $col_idx (0..$#$cols) {
1675 my $val = $datum->[$col_idx];
1676 my $sqla_bind = $colvalues{ $cols->[$col_idx] };
1677 my $is_literal_sql = (ref $sqla_bind) eq 'SCALAR';
1679 if ($is_literal_sql) {
1681 $bad_slice->('bind found where literal SQL expected', $col_idx, $datum_idx);
1683 elsif ((my $reftype = ref $val) ne 'SCALAR') {
1684 $bad_slice->("$reftype reference found where literal SQL expected",
1685 $col_idx, $datum_idx);
1687 elsif ($$val ne $$sqla_bind){
1688 $bad_slice->("inconsistent literal SQL value, expecting: '$$sqla_bind'",
1689 $col_idx, $datum_idx);
1692 elsif (my $reftype = ref $val) {
1694 if (overload::Method($val, '""')) {
1695 $datum->[$col_idx] = "".$val;
1698 $bad_slice->("$reftype reference found where bind expected",
1699 $col_idx, $datum_idx);
1705 my ($sql, $bind) = $self->_prep_for_execute (
1706 'insert', undef, $source, [\%colvalues]
1710 my $empty_bind = 1 if (not @bind) &&
1711 (grep { ref $_ eq 'SCALAR' } values %colvalues) == @$cols;
1713 if ((not @bind) && (not $empty_bind)) {
1714 $self->throw_exception(
1715 'Cannot insert_bulk without support for placeholders'
1719 # neither _execute_array, nor _execute_inserts_with_no_binds are
1720 # atomic (even if _execute _array is a single call). Thus a safety
1722 my $guard = $self->txn_scope_guard;
1724 $self->_query_start( $sql, [ dummy => '__BULK_INSERT__' ] );
1725 my $sth = $self->sth($sql);
1728 # bind_param_array doesn't work if there are no binds
1729 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
1732 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1733 $self->_execute_array( $source, $sth, \@bind, $cols, $data );
1737 $self->_query_end( $sql, [ dummy => '__BULK_INSERT__' ] );
1741 return (wantarray ? ($rv, $sth, @bind) : $rv);
1744 sub _execute_array {
1745 my ($self, $source, $sth, $bind, $cols, $data, @extra) = @_;
1747 ## This must be an arrayref, else nothing works!
1748 my $tuple_status = [];
1750 ## Get the bind_attributes, if any exist
1751 my $bind_attributes = $self->source_bind_attributes($source);
1753 ## Bind the values and execute
1754 my $placeholder_index = 1;
1756 foreach my $bound (@$bind) {
1758 my $attributes = {};
1759 my ($column_name, $data_index) = @$bound;
1761 if( $bind_attributes ) {
1762 $attributes = $bind_attributes->{$column_name}
1763 if defined $bind_attributes->{$column_name};
1766 my @data = map { $_->[$data_index] } @$data;
1768 $sth->bind_param_array(
1771 (%$attributes ? $attributes : ()),
1773 $placeholder_index++;
1778 $rv = $self->_dbh_execute_array($sth, $tuple_status, @extra);
1784 # Statement must finish even if there was an exception.
1789 $err = shift unless defined $err
1794 if (! defined $err and $sth->err);
1798 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1800 $self->throw_exception("Unexpected populate error: $err")
1801 if ($i > $#$tuple_status);
1803 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1804 ($tuple_status->[$i][1] || $err),
1805 Dumper { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) },
1812 sub _dbh_execute_array {
1813 my ($self, $sth, $tuple_status, @extra) = @_;
1815 return $sth->execute_array({ArrayTupleStatus => $tuple_status});
1818 sub _dbh_execute_inserts_with_no_binds {
1819 my ($self, $sth, $count) = @_;
1823 my $dbh = $self->_get_dbh;
1824 local $dbh->{RaiseError} = 1;
1825 local $dbh->{PrintError} = 0;
1827 $sth->execute foreach 1..$count;
1833 # Make sure statement is finished even if there was an exception.
1838 $err = shift unless defined $err;
1842 $self->throw_exception($err) if defined $err;
1848 my ($self, $source, @args) = @_;
1850 my $bind_attrs = $self->source_bind_attributes($source);
1852 return $self->_execute('update' => [], $source, $bind_attrs, @args);
1857 my ($self, $source, @args) = @_;
1859 my $bind_attrs = $self->source_bind_attributes($source);
1861 return $self->_execute('delete' => [], $source, $bind_attrs, @args);
1864 # We were sent here because the $rs contains a complex search
1865 # which will require a subquery to select the correct rows
1866 # (i.e. joined or limited resultsets, or non-introspectable conditions)
1868 # Generating a single PK column subquery is trivial and supported
1869 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1870 # Look at _multipk_update_delete()
1871 sub _subq_update_delete {
1873 my ($rs, $op, $values) = @_;
1875 my $rsrc = $rs->result_source;
1877 # quick check if we got a sane rs on our hands
1878 my @pcols = $rsrc->_pri_cols;
1880 my $sel = $rs->_resolved_attrs->{select};
1881 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1884 join ("\x00", map { join '.', $rs->{attrs}{alias}, $_ } sort @pcols)
1886 join ("\x00", sort @$sel )
1888 $self->throw_exception (
1889 '_subq_update_delete can not be called on resultsets selecting columns other than the primary keys'
1896 $op eq 'update' ? $values : (),
1897 { $pcols[0] => { -in => $rs->as_query } },
1902 return $self->_multipk_update_delete (@_);
1906 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1907 # resultset update/delete involving subqueries. So by default resort
1908 # to simple (and inefficient) delete_all style per-row opearations,
1909 # while allowing specific storages to override this with a faster
1912 sub _multipk_update_delete {
1913 return shift->_per_row_update_delete (@_);
1916 # This is the default loop used to delete/update rows for multi PK
1917 # resultsets, and used by mysql exclusively (because it can't do anything
1920 # We do not use $row->$op style queries, because resultset update/delete
1921 # is not expected to cascade (this is what delete_all/update_all is for).
1923 # There should be no race conditions as the entire operation is rolled
1926 sub _per_row_update_delete {
1928 my ($rs, $op, $values) = @_;
1930 my $rsrc = $rs->result_source;
1931 my @pcols = $rsrc->_pri_cols;
1933 my $guard = $self->txn_scope_guard;
1935 # emulate the return value of $sth->execute for non-selects
1936 my $row_cnt = '0E0';
1938 my $subrs_cur = $rs->cursor;
1939 my @all_pk = $subrs_cur->all;
1940 for my $pks ( @all_pk) {
1943 for my $i (0.. $#pcols) {
1944 $cond->{$pcols[$i]} = $pks->[$i];
1949 $op eq 'update' ? $values : (),
1963 $self->_execute($self->_select_args(@_));
1966 sub _select_args_to_query {
1969 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $rs_attrs, $rows, $offset)
1970 # = $self->_select_args($ident, $select, $cond, $attrs);
1971 my ($op, $bind, $ident, $bind_attrs, @args) =
1972 $self->_select_args(@_);
1974 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
1975 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1976 $prepared_bind ||= [];
1979 ? ($sql, $prepared_bind, $bind_attrs)
1980 : \[ "($sql)", @$prepared_bind ]
1985 my ($self, $ident, $select, $where, $attrs) = @_;
1987 my $sql_maker = $self->sql_maker;
1988 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1995 $rs_alias && $alias2source->{$rs_alias}
1996 ? ( _rsroot_source_handle => $alias2source->{$rs_alias}->handle )
2001 # calculate bind_attrs before possible $ident mangling
2002 my $bind_attrs = {};
2003 for my $alias (keys %$alias2source) {
2004 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
2005 for my $col (keys %$bindtypes) {
2007 my $fqcn = join ('.', $alias, $col);
2008 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
2010 # Unqialified column names are nice, but at the same time can be
2011 # rather ambiguous. What we do here is basically go along with
2012 # the loop, adding an unqualified column slot to $bind_attrs,
2013 # alongside the fully qualified name. As soon as we encounter
2014 # another column by that name (which would imply another table)
2015 # we unset the unqualified slot and never add any info to it
2016 # to avoid erroneous type binding. If this happens the users
2017 # only choice will be to fully qualify his column name
2019 if (exists $bind_attrs->{$col}) {
2020 $bind_attrs->{$col} = {};
2023 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
2028 # Sanity check the attributes (SQLMaker does it too, but
2029 # in case of a software_limit we'll never reach there)
2030 if (defined $attrs->{offset}) {
2031 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
2032 if ( $attrs->{offset} =~ /\D/ or $attrs->{offset} < 0 );
2034 $attrs->{offset} ||= 0;
2036 if (defined $attrs->{rows}) {
2037 $self->throw_exception("The rows attribute must be a positive integer if present")
2038 if ( $attrs->{rows} =~ /\D/ or $attrs->{rows} <= 0 );
2040 elsif ($attrs->{offset}) {
2041 # MySQL actually recommends this approach. I cringe.
2042 $attrs->{rows} = $sql_maker->__max_int;
2047 # see if we need to tear the prefetch apart otherwise delegate the limiting to the
2048 # storage, unless software limit was requested
2051 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
2053 # grouped prefetch (to satisfy group_by == select)
2054 ( $attrs->{group_by}
2056 @{$attrs->{group_by}}
2058 $attrs->{_prefetch_select}
2060 @{$attrs->{_prefetch_select}}
2063 ($ident, $select, $where, $attrs)
2064 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
2066 elsif (! $attrs->{software_limit} ) {
2067 push @limit, $attrs->{rows}, $attrs->{offset};
2070 # try to simplify the joinmap further (prune unreferenced type-single joins)
2071 $ident = $self->_prune_unused_joins ($ident, $select, $where, $attrs);
2074 # This would be the point to deflate anything found in $where
2075 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2076 # expect a row object. And all we have is a resultsource (it is trivial
2077 # to extract deflator coderefs via $alias2source above).
2079 # I don't see a way forward other than changing the way deflators are
2080 # invoked, and that's just bad...
2083 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $attrs, @limit);
2086 # Returns a counting SELECT for a simple count
2087 # query. Abstracted so that a storage could override
2088 # this to { count => 'firstcol' } or whatever makes
2089 # sense as a performance optimization
2091 #my ($self, $source, $rs_attrs) = @_;
2092 return { count => '*' };
2096 sub source_bind_attributes {
2097 my ($self, $source) = @_;
2099 my $bind_attributes;
2100 foreach my $column ($source->columns) {
2102 my $data_type = $source->column_info($column)->{data_type} || '';
2103 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
2107 return $bind_attributes;
2114 =item Arguments: $ident, $select, $condition, $attrs
2118 Handle a SQL select statement.
2124 my ($ident, $select, $condition, $attrs) = @_;
2125 return $self->cursor_class->new($self, \@_, $attrs);
2130 my ($rv, $sth, @bind) = $self->_select(@_);
2131 my @row = $sth->fetchrow_array;
2132 my @nextrow = $sth->fetchrow_array if @row;
2133 if(@row && @nextrow) {
2134 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2136 # Need to call finish() to work round broken DBDs
2141 =head2 sql_limit_dialect
2143 This is an accessor for the default SQL limit dialect used by a particular
2144 storage driver. Can be overriden by supplying an explicit L</limit_dialect>
2145 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2146 see L<DBIx::Class::SQLMaker::LimitDialects>.
2152 =item Arguments: $sql
2156 Returns a L<DBI> sth (statement handle) for the supplied SQL.
2161 my ($self, $dbh, $sql) = @_;
2163 # 3 is the if_active parameter which avoids active sth re-use
2164 my $sth = $self->disable_sth_caching
2165 ? $dbh->prepare($sql)
2166 : $dbh->prepare_cached($sql, {}, 3);
2168 # XXX You would think RaiseError would make this impossible,
2169 # but apparently that's not true :(
2170 $self->throw_exception($dbh->errstr) if !$sth;
2176 my ($self, $sql) = @_;
2177 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2180 sub _dbh_columns_info_for {
2181 my ($self, $dbh, $table) = @_;
2183 if ($dbh->can('column_info')) {
2187 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2188 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2190 while ( my $info = $sth->fetchrow_hashref() ){
2192 $column_info{data_type} = $info->{TYPE_NAME};
2193 $column_info{size} = $info->{COLUMN_SIZE};
2194 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2195 $column_info{default_value} = $info->{COLUMN_DEF};
2196 my $col_name = $info->{COLUMN_NAME};
2197 $col_name =~ s/^\"(.*)\"$/$1/;
2199 $result{$col_name} = \%column_info;
2204 return \%result if !$caught && scalar keys %result;
2208 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2210 my @columns = @{$sth->{NAME_lc}};
2211 for my $i ( 0 .. $#columns ){
2213 $column_info{data_type} = $sth->{TYPE}->[$i];
2214 $column_info{size} = $sth->{PRECISION}->[$i];
2215 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2217 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2218 $column_info{data_type} = $1;
2219 $column_info{size} = $2;
2222 $result{$columns[$i]} = \%column_info;
2226 foreach my $col (keys %result) {
2227 my $colinfo = $result{$col};
2228 my $type_num = $colinfo->{data_type};
2230 if(defined $type_num && $dbh->can('type_info')) {
2231 my $type_info = $dbh->type_info($type_num);
2232 $type_name = $type_info->{TYPE_NAME} if $type_info;
2233 $colinfo->{data_type} = $type_name if $type_name;
2240 sub columns_info_for {
2241 my ($self, $table) = @_;
2242 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2245 =head2 last_insert_id
2247 Return the row id of the last insert.
2251 sub _dbh_last_insert_id {
2252 my ($self, $dbh, $source, $col) = @_;
2254 my $id = try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2256 return $id if defined $id;
2258 my $class = ref $self;
2259 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2262 sub last_insert_id {
2264 $self->_dbh_last_insert_id ($self->_dbh, @_);
2267 =head2 _native_data_type
2271 =item Arguments: $type_name
2275 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2276 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2277 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2279 The default implementation returns C<undef>, implement in your Storage driver if
2280 you need this functionality.
2282 Should map types from other databases to the native RDBMS type, for example
2283 C<VARCHAR2> to C<VARCHAR>.
2285 Types with modifiers should map to the underlying data type. For example,
2286 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2288 Composite types should map to the container type, for example
2289 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2293 sub _native_data_type {
2294 #my ($self, $data_type) = @_;
2298 # Check if placeholders are supported at all
2299 sub _determine_supports_placeholders {
2301 my $dbh = $self->_get_dbh;
2303 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2304 # but it is inaccurate more often than not
2306 local $dbh->{PrintError} = 0;
2307 local $dbh->{RaiseError} = 1;
2308 $dbh->do('select ?', {}, 1);
2316 # Check if placeholders bound to non-string types throw exceptions
2318 sub _determine_supports_typeless_placeholders {
2320 my $dbh = $self->_get_dbh;
2323 local $dbh->{PrintError} = 0;
2324 local $dbh->{RaiseError} = 1;
2325 # this specifically tests a bind that is NOT a string
2326 $dbh->do('select 1 where 1 = ?', {}, 1);
2336 Returns the database driver name.
2341 shift->_get_dbh->{Driver}->{Name};
2344 =head2 bind_attribute_by_data_type
2346 Given a datatype from column info, returns a database specific bind
2347 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2348 let the database planner just handle it.
2350 Generally only needed for special case column types, like bytea in postgres.
2354 sub bind_attribute_by_data_type {
2358 =head2 is_datatype_numeric
2360 Given a datatype from column_info, returns a boolean value indicating if
2361 the current RDBMS considers it a numeric value. This controls how
2362 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2363 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2364 be performed instead of the usual C<eq>.
2368 sub is_datatype_numeric {
2369 my ($self, $dt) = @_;
2371 return 0 unless $dt;
2373 return $dt =~ /^ (?:
2374 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2379 =head2 create_ddl_dir
2383 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2387 Creates a SQL file based on the Schema, for each of the specified
2388 database engines in C<\@databases> in the given directory.
2389 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2391 Given a previous version number, this will also create a file containing
2392 the ALTER TABLE statements to transform the previous schema into the
2393 current one. Note that these statements may contain C<DROP TABLE> or
2394 C<DROP COLUMN> statements that can potentially destroy data.
2396 The file names are created using the C<ddl_filename> method below, please
2397 override this method in your schema if you would like a different file
2398 name format. For the ALTER file, the same format is used, replacing
2399 $version in the name with "$preversion-$version".
2401 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2402 The most common value for this would be C<< { add_drop_table => 1 } >>
2403 to have the SQL produced include a C<DROP TABLE> statement for each table
2404 created. For quoting purposes supply C<quote_table_names> and
2405 C<quote_field_names>.
2407 If no arguments are passed, then the following default values are assumed:
2411 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2413 =item version - $schema->schema_version
2415 =item directory - './'
2417 =item preversion - <none>
2421 By default, C<\%sqlt_args> will have
2423 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2425 merged with the hash passed in. To disable any of those features, pass in a
2426 hashref like the following
2428 { ignore_constraint_names => 0, # ... other options }
2431 WARNING: You are strongly advised to check all SQL files created, before applying
2436 sub create_ddl_dir {
2437 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2440 carp "No directory given, using ./\n";
2445 make_path ("$dir") # make_path does not like objects (i.e. Path::Class::Dir)
2447 $self->throw_exception(
2448 "Failed to create '$dir': " . ($! || $@ || 'error unknow')
2452 $self->throw_exception ("Directory '$dir' does not exist\n") unless(-d $dir);
2454 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2455 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2457 my $schema_version = $schema->schema_version || '1.x';
2458 $version ||= $schema_version;
2461 add_drop_table => 1,
2462 ignore_constraint_names => 1,
2463 ignore_index_names => 1,
2467 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy')) {
2468 $self->throw_exception("Can't create a ddl file without " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2471 my $sqlt = SQL::Translator->new( $sqltargs );
2473 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2474 my $sqlt_schema = $sqlt->translate({ data => $schema })
2475 or $self->throw_exception ($sqlt->error);
2477 foreach my $db (@$databases) {
2479 $sqlt->{schema} = $sqlt_schema;
2480 $sqlt->producer($db);
2483 my $filename = $schema->ddl_filename($db, $version, $dir);
2484 if (-e $filename && ($version eq $schema_version )) {
2485 # if we are dumping the current version, overwrite the DDL
2486 carp "Overwriting existing DDL file - $filename";
2490 my $output = $sqlt->translate;
2492 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2495 if(!open($file, ">$filename")) {
2496 $self->throw_exception("Can't open $filename for writing ($!)");
2499 print $file $output;
2502 next unless ($preversion);
2504 require SQL::Translator::Diff;
2506 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2507 if(!-e $prefilename) {
2508 carp("No previous schema file found ($prefilename)");
2512 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2514 carp("Overwriting existing diff file - $difffile");
2520 my $t = SQL::Translator->new($sqltargs);
2525 or $self->throw_exception ($t->error);
2527 my $out = $t->translate( $prefilename )
2528 or $self->throw_exception ($t->error);
2530 $source_schema = $t->schema;
2532 $source_schema->name( $prefilename )
2533 unless ( $source_schema->name );
2536 # The "new" style of producers have sane normalization and can support
2537 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2538 # And we have to diff parsed SQL against parsed SQL.
2539 my $dest_schema = $sqlt_schema;
2541 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2542 my $t = SQL::Translator->new($sqltargs);
2547 or $self->throw_exception ($t->error);
2549 my $out = $t->translate( $filename )
2550 or $self->throw_exception ($t->error);
2552 $dest_schema = $t->schema;
2554 $dest_schema->name( $filename )
2555 unless $dest_schema->name;
2558 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2562 if(!open $file, ">$difffile") {
2563 $self->throw_exception("Can't write to $difffile ($!)");
2571 =head2 deployment_statements
2575 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2579 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2581 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2582 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2584 C<$directory> is used to return statements from files in a previously created
2585 L</create_ddl_dir> directory and is optional. The filenames are constructed
2586 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2588 If no C<$directory> is specified then the statements are constructed on the
2589 fly using L<SQL::Translator> and C<$version> is ignored.
2591 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2595 sub deployment_statements {
2596 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2597 $type ||= $self->sqlt_type;
2598 $version ||= $schema->schema_version || '1.x';
2600 my $filename = $schema->ddl_filename($type, $version, $dir);
2604 open($file, "<$filename")
2605 or $self->throw_exception("Can't open $filename ($!)");
2608 return join('', @rows);
2611 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy') ) {
2612 $self->throw_exception("Can't deploy without a ddl_dir or " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2615 # sources needs to be a parser arg, but for simplicty allow at top level
2617 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2618 if exists $sqltargs->{sources};
2620 my $tr = SQL::Translator->new(
2621 producer => "SQL::Translator::Producer::${type}",
2623 parser => 'SQL::Translator::Parser::DBIx::Class',
2630 @ret = $tr->translate;
2633 $ret[0] = $tr->translate;
2636 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
2637 unless (@ret && defined $ret[0]);
2639 return $wa ? @ret : $ret[0];
2643 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2646 return if($line =~ /^--/);
2648 # next if($line =~ /^DROP/m);
2649 return if($line =~ /^BEGIN TRANSACTION/m);
2650 return if($line =~ /^COMMIT/m);
2651 return if $line =~ /^\s+$/; # skip whitespace only
2652 $self->_query_start($line);
2654 # do a dbh_do cycle here, as we need some error checking in
2655 # place (even though we will ignore errors)
2656 $self->dbh_do (sub { $_[1]->do($line) });
2658 carp qq{$_ (running "${line}")};
2660 $self->_query_end($line);
2662 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2663 if (@statements > 1) {
2664 foreach my $statement (@statements) {
2665 $deploy->( $statement );
2668 elsif (@statements == 1) {
2669 foreach my $line ( split(";\n", $statements[0])) {
2675 =head2 datetime_parser
2677 Returns the datetime parser class
2681 sub datetime_parser {
2683 return $self->{datetime_parser} ||= do {
2684 $self->build_datetime_parser(@_);
2688 =head2 datetime_parser_type
2690 Defines (returns) the datetime parser class - currently hardwired to
2691 L<DateTime::Format::MySQL>
2695 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2697 =head2 build_datetime_parser
2699 See L</datetime_parser>
2703 sub build_datetime_parser {
2705 my $type = $self->datetime_parser_type(@_);
2706 $self->ensure_class_loaded ($type);
2711 =head2 is_replicating
2713 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2714 replicate from a master database. Default is undef, which is the result
2715 returned by databases that don't support replication.
2719 sub is_replicating {
2724 =head2 lag_behind_master
2726 Returns a number that represents a certain amount of lag behind a master db
2727 when a given storage is replicating. The number is database dependent, but
2728 starts at zero and increases with the amount of lag. Default in undef
2732 sub lag_behind_master {
2736 =head2 relname_to_table_alias
2740 =item Arguments: $relname, $join_count
2744 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
2747 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
2748 way these aliases are named.
2750 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
2751 otherwise C<"$relname">.
2755 sub relname_to_table_alias {
2756 my ($self, $relname, $join_count) = @_;
2758 my $alias = ($join_count && $join_count > 1 ?
2759 join('_', $relname, $join_count) : $relname);
2768 =head2 DBIx::Class and AutoCommit
2770 DBIx::Class can do some wonderful magic with handling exceptions,
2771 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2772 (the default) combined with C<txn_do> for transaction support.
2774 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2775 in an assumed transaction between commits, and you're telling us you'd
2776 like to manage that manually. A lot of the magic protections offered by
2777 this module will go away. We can't protect you from exceptions due to database
2778 disconnects because we don't know anything about how to restart your
2779 transactions. You're on your own for handling all sorts of exceptional
2780 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2786 Matt S. Trout <mst@shadowcatsystems.co.uk>
2788 Andy Grundman <andy@hybridized.org>
2792 You may distribute this code under the same terms as Perl itself.