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
86 for my $meth (@rdbms_specific_methods) {
88 my $orig = __PACKAGE__->can ($meth)
89 or die "$meth is not a ::Storage::DBI method!";
92 no warnings qw/redefine/;
93 *{__PACKAGE__ ."::$meth"} = subname $meth => sub {
94 if (not $_[0]->_driver_determined and not $_[0]->{_in_determine_driver}) {
95 $_[0]->_determine_driver;
97 # This for some reason crashes and burns on perl 5.8.1
98 # IFF the method ends up throwing an exception
99 #goto $_[0]->can ($meth);
101 my $cref = $_[0]->can ($meth);
111 DBIx::Class::Storage::DBI - DBI storage handler
115 my $schema = MySchema->connect('dbi:SQLite:my.db');
117 $schema->storage->debug(1);
119 my @stuff = $schema->storage->dbh_do(
121 my ($storage, $dbh, @args) = @_;
122 $dbh->do("DROP TABLE authors");
127 $schema->resultset('Book')->search({
128 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
133 This class represents the connection to an RDBMS via L<DBI>. See
134 L<DBIx::Class::Storage> for general information. This pod only
135 documents DBI-specific methods and behaviors.
142 my $new = shift->next::method(@_);
144 $new->transaction_depth(0);
145 $new->_sql_maker_opts({});
146 $new->_dbh_details({});
147 $new->{savepoints} = [];
148 $new->{_in_dbh_do} = 0;
149 $new->{_dbh_gen} = 0;
151 # read below to see what this does
152 $new->_arm_global_destructor;
157 # This is hack to work around perl shooting stuff in random
158 # order on exit(). If we do not walk the remaining storage
159 # objects in an END block, there is a *small but real* chance
160 # of a fork()ed child to kill the parent's shared DBI handle,
161 # *before perl reaches the DESTROY in this package*
162 # Yes, it is ugly and effective.
164 my %seek_and_destroy;
166 sub _arm_global_destructor {
168 my $key = Scalar::Util::refaddr ($self);
169 $seek_and_destroy{$key} = $self;
170 Scalar::Util::weaken ($seek_and_destroy{$key});
174 local $?; # just in case the DBI destructor changes it somehow
176 # destroy just the object if not native to this process/thread
177 $_->_preserve_foreign_dbh for (grep
179 values %seek_and_destroy
187 # destroy just the object if not native to this process/thread
188 $self->_preserve_foreign_dbh;
190 # some databases need this to stop spewing warnings
191 if (my $dbh = $self->_dbh) {
193 %{ $dbh->{CachedKids} } = ();
201 sub _preserve_foreign_dbh {
204 return unless $self->_dbh;
208 return unless $self->_dbh;
214 # handle pid changes correctly - do not destroy parent's connection
218 return if ( defined $self->_conn_pid and $self->_conn_pid == $$ );
220 $self->_dbh->{InactiveDestroy} = 1;
227 # very similar to above, but seems to FAIL if I set InactiveDestroy
231 if ( ! defined $self->_conn_tid ) {
234 elsif ( $self->_conn_tid == threads->tid ) {
235 return; # same thread
238 #$self->_dbh->{InactiveDestroy} = 1; # why does t/51threads.t fail...?
248 This method is normally called by L<DBIx::Class::Schema/connection>, which
249 encapsulates its argument list in an arrayref before passing them here.
251 The argument list may contain:
257 The same 4-element argument set one would normally pass to
258 L<DBI/connect>, optionally followed by
259 L<extra attributes|/DBIx::Class specific connection attributes>
260 recognized by DBIx::Class:
262 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
266 A single code reference which returns a connected
267 L<DBI database handle|DBI/connect> optionally followed by
268 L<extra attributes|/DBIx::Class specific connection attributes> recognized
271 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
275 A single hashref with all the attributes and the dsn/user/password
278 $connect_info_args = [{
286 $connect_info_args = [{
287 dbh_maker => sub { DBI->connect (...) },
292 This is particularly useful for L<Catalyst> based applications, allowing the
293 following config (L<Config::General> style):
298 dsn dbi:mysql:database=test
305 The C<dsn>/C<user>/C<password> combination can be substituted by the
306 C<dbh_maker> key whose value is a coderef that returns a connected
307 L<DBI database handle|DBI/connect>
311 Please note that the L<DBI> docs recommend that you always explicitly
312 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
313 recommends that it be set to I<1>, and that you perform transactions
314 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
315 to I<1> if you do not do explicitly set it to zero. This is the default
316 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
318 =head3 DBIx::Class specific connection attributes
320 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
321 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
322 the following connection options. These options can be mixed in with your other
323 L<DBI> connection attributes, or placed in a separate hashref
324 (C<\%extra_attributes>) as shown above.
326 Every time C<connect_info> is invoked, any previous settings for
327 these options will be cleared before setting the new ones, regardless of
328 whether any options are specified in the new C<connect_info>.
335 Specifies things to do immediately after connecting or re-connecting to
336 the database. Its value may contain:
342 This contains one SQL statement to execute.
344 =item an array reference
346 This contains SQL statements to execute in order. Each element contains
347 a string or a code reference that returns a string.
349 =item a code reference
351 This contains some code to execute. Unlike code references within an
352 array reference, its return value is ignored.
356 =item on_disconnect_do
358 Takes arguments in the same form as L</on_connect_do> and executes them
359 immediately before disconnecting from the database.
361 Note, this only runs if you explicitly call L</disconnect> on the
364 =item on_connect_call
366 A more generalized form of L</on_connect_do> that calls the specified
367 C<connect_call_METHOD> methods in your storage driver.
369 on_connect_do => 'select 1'
373 on_connect_call => [ [ do_sql => 'select 1' ] ]
375 Its values may contain:
381 Will call the C<connect_call_METHOD> method.
383 =item a code reference
385 Will execute C<< $code->($storage) >>
387 =item an array reference
389 Each value can be a method name or code reference.
391 =item an array of arrays
393 For each array, the first item is taken to be the C<connect_call_> method name
394 or code reference, and the rest are parameters to it.
398 Some predefined storage methods you may use:
404 Executes a SQL string or a code reference that returns a SQL string. This is
405 what L</on_connect_do> and L</on_disconnect_do> use.
413 Will execute the scalar as SQL.
417 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
418 attributes hashref and bind values.
420 =item a code reference
422 Will execute C<< $code->($storage) >> and execute the return array refs as
429 Execute any statements necessary to initialize the database session to return
430 and accept datetime/timestamp values used with
431 L<DBIx::Class::InflateColumn::DateTime>.
433 Only necessary for some databases, see your specific storage driver for
434 implementation details.
438 =item on_disconnect_call
440 Takes arguments in the same form as L</on_connect_call> and executes them
441 immediately before disconnecting from the database.
443 Calls the C<disconnect_call_METHOD> methods as opposed to the
444 C<connect_call_METHOD> methods called by L</on_connect_call>.
446 Note, this only runs if you explicitly call L</disconnect> on the
449 =item disable_sth_caching
451 If set to a true value, this option will disable the caching of
452 statement handles via L<DBI/prepare_cached>.
456 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
457 default L</sql_limit_dialect> setting of the storage (if any). For a list
458 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
462 Specifies what characters to use to quote table and column names. If
463 you use this you will want to specify L</name_sep> as well.
465 C<quote_char> expects either a single character, in which case is it
466 is placed on either side of the table/column name, or an arrayref of length
467 2 in which case the table/column name is placed between the elements.
469 For example under MySQL you should use C<< quote_char => '`' >>, and for
470 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
474 This only needs to be used in conjunction with C<quote_char>, and is used to
475 specify the character that separates elements (schemas, tables, columns) from
476 each other. In most cases this is simply a C<.>.
478 The consequences of not supplying this value is that L<SQL::Abstract>
479 will assume DBIx::Class' uses of aliases to be complete column
480 names. The output will look like I<"me.name"> when it should actually
485 This Storage driver normally installs its own C<HandleError>, sets
486 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
487 all database handles, including those supplied by a coderef. It does this
488 so that it can have consistent and useful error behavior.
490 If you set this option to a true value, Storage will not do its usual
491 modifications to the database handle's attributes, and instead relies on
492 the settings in your connect_info DBI options (or the values you set in
493 your connection coderef, in the case that you are connecting via coderef).
495 Note that your custom settings can cause Storage to malfunction,
496 especially if you set a C<HandleError> handler that suppresses exceptions
497 and/or disable C<RaiseError>.
501 If this option is true, L<DBIx::Class> will use savepoints when nesting
502 transactions, making it possible to recover from failure in the inner
503 transaction without having to abort all outer transactions.
507 Use this argument to supply a cursor class other than the default
508 L<DBIx::Class::Storage::DBI::Cursor>.
512 Some real-life examples of arguments to L</connect_info> and
513 L<DBIx::Class::Schema/connect>
515 # Simple SQLite connection
516 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
519 ->connect_info([ sub { DBI->connect(...) } ]);
521 # Connect via subref in hashref
523 dbh_maker => sub { DBI->connect(...) },
524 on_connect_do => 'alter session ...',
527 # A bit more complicated
534 { quote_char => q{"}, name_sep => q{.} },
538 # Equivalent to the previous example
544 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
548 # Same, but with hashref as argument
549 # See parse_connect_info for explanation
552 dsn => 'dbi:Pg:dbname=foo',
554 password => 'my_pg_password',
561 # Subref + DBIx::Class-specific connection options
564 sub { DBI->connect(...) },
568 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
569 disable_sth_caching => 1,
579 my ($self, $info) = @_;
581 return $self->_connect_info if !$info;
583 $self->_connect_info($info); # copy for _connect_info
585 $info = $self->_normalize_connect_info($info)
586 if ref $info eq 'ARRAY';
588 for my $storage_opt (keys %{ $info->{storage_options} }) {
589 my $value = $info->{storage_options}{$storage_opt};
591 $self->$storage_opt($value);
594 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
595 # the new set of options
596 $self->_sql_maker(undef);
597 $self->_sql_maker_opts({});
599 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
600 my $value = $info->{sql_maker_options}{$sql_maker_opt};
602 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
606 %{ $self->_default_dbi_connect_attributes || {} },
607 %{ $info->{attributes} || {} },
610 my @args = @{ $info->{arguments} };
612 $self->_dbi_connect_info([@args,
613 %attrs && !(ref $args[0] eq 'CODE') ? \%attrs : ()]);
616 # save attributes them in a separate accessor so they are always
617 # introspectable, even in case of a CODE $dbhmaker
618 $self->_dbic_connect_attributes (\%attrs);
620 return $self->_connect_info;
623 sub _normalize_connect_info {
624 my ($self, $info_arg) = @_;
627 my @args = @$info_arg; # take a shallow copy for further mutilation
629 # combine/pre-parse arguments depending on invocation style
632 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
633 %attrs = %{ $args[1] || {} };
636 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
637 %attrs = %{$args[0]};
639 if (my $code = delete $attrs{dbh_maker}) {
642 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
645 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
646 . "to the result of 'dbh_maker'",
648 join (', ', map { "'$_'" } (@ignored) ),
653 @args = delete @attrs{qw/dsn user password/};
656 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
658 % { $args[3] || {} },
659 % { $args[4] || {} },
661 @args = @args[0,1,2];
664 $info{arguments} = \@args;
666 my @storage_opts = grep exists $attrs{$_},
667 @storage_options, 'cursor_class';
669 @{ $info{storage_options} }{@storage_opts} =
670 delete @attrs{@storage_opts} if @storage_opts;
672 my @sql_maker_opts = grep exists $attrs{$_},
673 qw/limit_dialect quote_char name_sep/;
675 @{ $info{sql_maker_options} }{@sql_maker_opts} =
676 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
678 $info{attributes} = \%attrs if %attrs;
683 sub _default_dbi_connect_attributes {
693 This method is deprecated in favour of setting via L</connect_info>.
697 =head2 on_disconnect_do
699 This method is deprecated in favour of setting via L</connect_info>.
703 sub _parse_connect_do {
704 my ($self, $type) = @_;
706 my $val = $self->$type;
707 return () if not defined $val;
712 push @res, [ 'do_sql', $val ];
713 } elsif (ref($val) eq 'CODE') {
715 } elsif (ref($val) eq 'ARRAY') {
716 push @res, map { [ 'do_sql', $_ ] } @$val;
718 $self->throw_exception("Invalid type for $type: ".ref($val));
726 Arguments: ($subref | $method_name), @extra_coderef_args?
728 Execute the given $subref or $method_name using the new exception-based
729 connection management.
731 The first two arguments will be the storage object that C<dbh_do> was called
732 on and a database handle to use. Any additional arguments will be passed
733 verbatim to the called subref as arguments 2 and onwards.
735 Using this (instead of $self->_dbh or $self->dbh) ensures correct
736 exception handling and reconnection (or failover in future subclasses).
738 Your subref should have no side-effects outside of the database, as
739 there is the potential for your subref to be partially double-executed
740 if the database connection was stale/dysfunctional.
744 my @stuff = $schema->storage->dbh_do(
746 my ($storage, $dbh, @cols) = @_;
747 my $cols = join(q{, }, @cols);
748 $dbh->selectrow_array("SELECT $cols FROM foo");
759 my $dbh = $self->_get_dbh;
761 return $self->$code($dbh, @_)
762 if ( $self->{_in_dbh_do} || $self->{transaction_depth} );
764 local $self->{_in_dbh_do} = 1;
766 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
769 $self->$code ($dbh, @$args);
771 $self->throw_exception($_) if $self->connected;
773 # We were not connected - reconnect and retry, but let any
774 # exception fall right through this time
775 carp "Retrying $code after catching disconnected exception: $_"
776 if $ENV{DBIC_DBIRETRY_DEBUG};
778 $self->_populate_dbh;
779 $self->$code($self->_dbh, @$args);
783 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
784 # It also informs dbh_do to bypass itself while under the direction of txn_do,
785 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
790 ref $coderef eq 'CODE' or $self->throw_exception
791 ('$coderef must be a CODE reference');
793 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
795 local $self->{_in_dbh_do} = 1;
798 my $want_array = wantarray;
804 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
812 @result = $coderef->(@$args);
814 elsif(defined $want_array) {
815 $result[0] = $coderef->(@$args);
825 if(! defined $exception) { return $want_array ? @result : $result[0] }
827 if($tried++ || $self->connected) {
828 my $rollback_exception;
829 try { $self->txn_rollback } catch { $rollback_exception = shift };
830 if(defined $rollback_exception) {
831 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
832 $self->throw_exception($exception) # propagate nested rollback
833 if $rollback_exception =~ /$exception_class/;
835 $self->throw_exception(
836 "Transaction aborted: ${exception}. "
837 . "Rollback failed: ${rollback_exception}"
840 $self->throw_exception($exception)
843 # We were not connected, and was first try - reconnect and retry
845 carp "Retrying $coderef after catching disconnected exception: $exception"
846 if $ENV{DBIC_DBIRETRY_DEBUG};
847 $self->_populate_dbh;
853 Our C<disconnect> method also performs a rollback first if the
854 database is not in C<AutoCommit> mode.
864 push @actions, ( $self->on_disconnect_call || () );
865 push @actions, $self->_parse_connect_do ('on_disconnect_do');
867 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
869 $self->_dbh_rollback unless $self->_dbh_autocommit;
871 %{ $self->_dbh->{CachedKids} } = ();
872 $self->_dbh->disconnect;
878 =head2 with_deferred_fk_checks
882 =item Arguments: C<$coderef>
884 =item Return Value: The return value of $coderef
888 Storage specific method to run the code ref with FK checks deferred or
889 in MySQL's case disabled entirely.
893 # Storage subclasses should override this
894 sub with_deferred_fk_checks {
895 my ($self, $sub) = @_;
903 =item Arguments: none
905 =item Return Value: 1|0
909 Verifies that the current database handle is active and ready to execute
910 an SQL statement (e.g. the connection did not get stale, server is still
911 answering, etc.) This method is used internally by L</dbh>.
917 return 0 unless $self->_seems_connected;
920 local $self->_dbh->{RaiseError} = 1;
925 sub _seems_connected {
928 $self->_preserve_foreign_dbh;
930 my $dbh = $self->_dbh
933 return $dbh->FETCH('Active');
939 my $dbh = $self->_dbh or return 0;
944 sub ensure_connected {
947 unless ($self->connected) {
948 $self->_populate_dbh;
954 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
955 is guaranteed to be healthy by implicitly calling L</connected>, and if
956 necessary performing a reconnection before returning. Keep in mind that this
957 is very B<expensive> on some database engines. Consider using L</dbh_do>
965 if (not $self->_dbh) {
966 $self->_populate_dbh;
968 $self->ensure_connected;
973 # this is the internal "get dbh or connect (don't check)" method
976 $self->_preserve_foreign_dbh;
977 $self->_populate_dbh unless $self->_dbh;
983 unless ($self->_sql_maker) {
984 my $sql_maker_class = $self->sql_maker_class;
985 $self->ensure_class_loaded ($sql_maker_class);
987 my %opts = %{$self->_sql_maker_opts||{}};
991 $self->sql_limit_dialect
994 my $s_class = (ref $self) || $self;
996 "Your storage class ($s_class) does not set sql_limit_dialect and you "
997 . 'have not supplied an explicit limit_dialect in your connection_info. '
998 . 'DBIC will attempt to use the GenericSubQ dialect, which works on most '
999 . 'databases but can be (and often is) painfully slow.'
1006 $self->_sql_maker($sql_maker_class->new(
1007 bindtype=>'columns',
1008 array_datatypes => 1,
1009 limit_dialect => $dialect,
1013 return $self->_sql_maker;
1016 # nothing to do by default
1023 my @info = @{$self->_dbi_connect_info || []};
1024 $self->_dbh(undef); # in case ->connected failed we might get sent here
1025 $self->_dbh_details({}); # reset everything we know
1027 $self->_dbh($self->_connect(@info));
1029 $self->_conn_pid($$);
1030 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
1032 $self->_determine_driver;
1034 # Always set the transaction depth on connect, since
1035 # there is no transaction in progress by definition
1036 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1038 $self->_run_connection_actions unless $self->{_in_determine_driver};
1041 sub _run_connection_actions {
1045 push @actions, ( $self->on_connect_call || () );
1046 push @actions, $self->_parse_connect_do ('on_connect_do');
1048 $self->_do_connection_actions(connect_call_ => $_) for @actions;
1053 sub set_use_dbms_capability {
1054 $_[0]->set_inherited ($_[1], $_[2]);
1057 sub get_use_dbms_capability {
1058 my ($self, $capname) = @_;
1060 my $use = $self->get_inherited ($capname);
1063 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1067 sub set_dbms_capability {
1068 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1071 sub get_dbms_capability {
1072 my ($self, $capname) = @_;
1074 my $cap = $self->_dbh_details->{capability}{$capname};
1076 unless (defined $cap) {
1077 if (my $meth = $self->can ("_determine$capname")) {
1078 $cap = $self->$meth ? 1 : 0;
1084 $self->set_dbms_capability ($capname, $cap);
1094 unless ($info = $self->_dbh_details->{info}) {
1098 my $server_version = try { $self->_get_server_version };
1100 if (defined $server_version) {
1101 $info->{dbms_version} = $server_version;
1103 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1104 my @verparts = split (/\./, $numeric_version);
1110 # consider only up to 3 version parts, iff not more than 3 digits
1112 while (@verparts && @use_parts < 3) {
1113 my $p = shift @verparts;
1115 push @use_parts, $p;
1117 push @use_parts, 0 while @use_parts < 3;
1119 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1123 $self->_dbh_details->{info} = $info;
1129 sub _get_server_version {
1130 shift->_get_dbh->get_info(18);
1133 sub _determine_driver {
1136 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1137 my $started_connected = 0;
1138 local $self->{_in_determine_driver} = 1;
1140 if (ref($self) eq __PACKAGE__) {
1142 if ($self->_dbh) { # we are connected
1143 $driver = $self->_dbh->{Driver}{Name};
1144 $started_connected = 1;
1146 # if connect_info is a CODEREF, we have no choice but to connect
1147 if (ref $self->_dbi_connect_info->[0] &&
1148 reftype $self->_dbi_connect_info->[0] eq 'CODE') {
1149 $self->_populate_dbh;
1150 $driver = $self->_dbh->{Driver}{Name};
1153 # try to use dsn to not require being connected, the driver may still
1154 # force a connection in _rebless to determine version
1155 # (dsn may not be supplied at all if all we do is make a mock-schema)
1156 my $dsn = $self->_dbi_connect_info->[0] || $ENV{DBI_DSN} || '';
1157 ($driver) = $dsn =~ /dbi:([^:]+):/i;
1158 $driver ||= $ENV{DBI_DRIVER};
1163 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1164 if ($self->load_optional_class($storage_class)) {
1165 mro::set_mro($storage_class, 'c3');
1166 bless $self, $storage_class;
1172 $self->_driver_determined(1);
1174 $self->_init; # run driver-specific initializations
1176 $self->_run_connection_actions
1177 if !$started_connected && defined $self->_dbh;
1181 sub _do_connection_actions {
1183 my $method_prefix = shift;
1186 if (not ref($call)) {
1187 my $method = $method_prefix . $call;
1189 } elsif (ref($call) eq 'CODE') {
1191 } elsif (ref($call) eq 'ARRAY') {
1192 if (ref($call->[0]) ne 'ARRAY') {
1193 $self->_do_connection_actions($method_prefix, $_) for @$call;
1195 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1198 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1204 sub connect_call_do_sql {
1206 $self->_do_query(@_);
1209 sub disconnect_call_do_sql {
1211 $self->_do_query(@_);
1214 # override in db-specific backend when necessary
1215 sub connect_call_datetime_setup { 1 }
1218 my ($self, $action) = @_;
1220 if (ref $action eq 'CODE') {
1221 $action = $action->($self);
1222 $self->_do_query($_) foreach @$action;
1225 # Most debuggers expect ($sql, @bind), so we need to exclude
1226 # the attribute hash which is the second argument to $dbh->do
1227 # furthermore the bind values are usually to be presented
1228 # as named arrayref pairs, so wrap those here too
1229 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1230 my $sql = shift @do_args;
1231 my $attrs = shift @do_args;
1232 my @bind = map { [ undef, $_ ] } @do_args;
1234 $self->_query_start($sql, @bind);
1235 $self->_get_dbh->do($sql, $attrs, @do_args);
1236 $self->_query_end($sql, @bind);
1243 my ($self, @info) = @_;
1245 $self->throw_exception("You failed to provide any connection info")
1248 my ($old_connect_via, $dbh);
1250 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
1251 $old_connect_via = $DBI::connect_via;
1252 $DBI::connect_via = 'connect';
1256 if(ref $info[0] eq 'CODE') {
1257 $dbh = $info[0]->();
1260 $dbh = DBI->connect(@info);
1267 unless ($self->unsafe) {
1269 # this odd anonymous coderef dereference is in fact really
1270 # necessary to avoid the unwanted effect described in perl5
1273 my $weak_self = $_[0];
1276 $_[1]->{HandleError} = sub {
1278 $weak_self->throw_exception("DBI Exception: $_[0]");
1281 # the handler may be invoked by something totally out of
1283 croak ("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1288 $dbh->{ShowErrorStatement} = 1;
1289 $dbh->{RaiseError} = 1;
1290 $dbh->{PrintError} = 0;
1294 $self->throw_exception("DBI Connection failed: $_")
1297 $DBI::connect_via = $old_connect_via if $old_connect_via;
1300 $self->_dbh_autocommit($dbh->{AutoCommit});
1305 my ($self, $name) = @_;
1307 $name = $self->_svp_generate_name
1308 unless defined $name;
1310 $self->throw_exception ("You can't use savepoints outside a transaction")
1311 if $self->{transaction_depth} == 0;
1313 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1314 unless $self->can('_svp_begin');
1316 push @{ $self->{savepoints} }, $name;
1318 $self->debugobj->svp_begin($name) if $self->debug;
1320 return $self->_svp_begin($name);
1324 my ($self, $name) = @_;
1326 $self->throw_exception ("You can't use savepoints outside a transaction")
1327 if $self->{transaction_depth} == 0;
1329 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1330 unless $self->can('_svp_release');
1332 if (defined $name) {
1333 $self->throw_exception ("Savepoint '$name' does not exist")
1334 unless grep { $_ eq $name } @{ $self->{savepoints} };
1336 # Dig through the stack until we find the one we are releasing. This keeps
1337 # the stack up to date.
1340 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1342 $name = pop @{ $self->{savepoints} };
1345 $self->debugobj->svp_release($name) if $self->debug;
1347 return $self->_svp_release($name);
1351 my ($self, $name) = @_;
1353 $self->throw_exception ("You can't use savepoints outside a transaction")
1354 if $self->{transaction_depth} == 0;
1356 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1357 unless $self->can('_svp_rollback');
1359 if (defined $name) {
1360 # If they passed us a name, verify that it exists in the stack
1361 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1362 $self->throw_exception("Savepoint '$name' does not exist!");
1365 # Dig through the stack until we find the one we are releasing. This keeps
1366 # the stack up to date.
1367 while(my $s = pop(@{ $self->{savepoints} })) {
1368 last if($s eq $name);
1370 # Add the savepoint back to the stack, as a rollback doesn't remove the
1371 # named savepoint, only everything after it.
1372 push(@{ $self->{savepoints} }, $name);
1374 # We'll assume they want to rollback to the last savepoint
1375 $name = $self->{savepoints}->[-1];
1378 $self->debugobj->svp_rollback($name) if $self->debug;
1380 return $self->_svp_rollback($name);
1383 sub _svp_generate_name {
1386 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1392 # this means we have not yet connected and do not know the AC status
1393 # (e.g. coderef $dbh)
1394 $self->ensure_connected if (! defined $self->_dbh_autocommit);
1396 if($self->{transaction_depth} == 0) {
1397 $self->debugobj->txn_begin()
1399 $self->_dbh_begin_work;
1401 elsif ($self->auto_savepoint) {
1404 $self->{transaction_depth}++;
1407 sub _dbh_begin_work {
1410 # if the user is utilizing txn_do - good for him, otherwise we need to
1411 # ensure that the $dbh is healthy on BEGIN.
1412 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1413 # will be replaced by a failure of begin_work itself (which will be
1414 # then retried on reconnect)
1415 if ($self->{_in_dbh_do}) {
1416 $self->_dbh->begin_work;
1418 $self->dbh_do(sub { $_[1]->begin_work });
1424 if ($self->{transaction_depth} == 1) {
1425 $self->debugobj->txn_commit()
1428 $self->{transaction_depth} = 0
1429 if $self->_dbh_autocommit;
1431 elsif($self->{transaction_depth} > 1) {
1432 $self->{transaction_depth}--;
1434 if $self->auto_savepoint;
1440 my $dbh = $self->_dbh
1441 or $self->throw_exception('cannot COMMIT on a disconnected handle');
1447 my $dbh = $self->_dbh;
1449 if ($self->{transaction_depth} == 1) {
1450 $self->debugobj->txn_rollback()
1452 $self->{transaction_depth} = 0
1453 if $self->_dbh_autocommit;
1454 $self->_dbh_rollback;
1456 elsif($self->{transaction_depth} > 1) {
1457 $self->{transaction_depth}--;
1458 if ($self->auto_savepoint) {
1459 $self->svp_rollback;
1464 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1468 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1470 if ($_ !~ /$exception_class/) {
1471 # ensure that a failed rollback resets the transaction depth
1472 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1475 $self->throw_exception($_)
1481 my $dbh = $self->_dbh
1482 or $self->throw_exception('cannot ROLLBACK on a disconnected handle');
1486 # This used to be the top-half of _execute. It was split out to make it
1487 # easier to override in NoBindVars without duping the rest. It takes up
1488 # all of _execute's args, and emits $sql, @bind.
1489 sub _prep_for_execute {
1490 my ($self, $op, $extra_bind, $ident, $args) = @_;
1492 if( blessed $ident && $ident->isa("DBIx::Class::ResultSource") ) {
1493 $ident = $ident->from();
1496 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1499 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1501 return ($sql, \@bind);
1505 sub _fix_bind_params {
1506 my ($self, @bind) = @_;
1508 ### Turn @bind from something like this:
1509 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1511 ### ( "'1'", "'1'", "'3'" )
1514 if ( defined( $_ && $_->[1] ) ) {
1515 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1522 my ( $self, $sql, @bind ) = @_;
1524 if ( $self->debug ) {
1525 @bind = $self->_fix_bind_params(@bind);
1527 $self->debugobj->query_start( $sql, @bind );
1532 my ( $self, $sql, @bind ) = @_;
1534 if ( $self->debug ) {
1535 @bind = $self->_fix_bind_params(@bind);
1536 $self->debugobj->query_end( $sql, @bind );
1541 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1543 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1545 $self->_query_start( $sql, @$bind );
1547 my $sth = $self->sth($sql,$op);
1549 my $placeholder_index = 1;
1551 foreach my $bound (@$bind) {
1552 my $attributes = {};
1553 my($column_name, @data) = @$bound;
1555 if ($bind_attributes) {
1556 $attributes = $bind_attributes->{$column_name}
1557 if defined $bind_attributes->{$column_name};
1560 foreach my $data (@data) {
1561 my $ref = ref $data;
1562 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1564 $sth->bind_param($placeholder_index, $data, $attributes);
1565 $placeholder_index++;
1569 # Can this fail without throwing an exception anyways???
1570 my $rv = $sth->execute();
1571 $self->throw_exception(
1572 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1575 $self->_query_end( $sql, @$bind );
1577 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1582 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1585 sub _prefetch_insert_auto_nextvals {
1586 my ($self, $source, $to_insert) = @_;
1590 foreach my $col ( $source->columns ) {
1591 if ( !defined $to_insert->{$col} ) {
1592 my $col_info = $source->column_info($col);
1594 if ( $col_info->{auto_nextval} ) {
1595 $upd->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1597 $col_info->{sequence} ||=
1598 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1609 my ($source, $to_insert, $opts) = @_;
1611 my $updated_cols = $self->_prefetch_insert_auto_nextvals (@_);
1613 my $bind_attributes = $self->source_bind_attributes($source);
1615 my ($rv, $sth) = $self->_execute('insert' => [], $source, $bind_attributes, $to_insert, $opts);
1617 if ($opts->{returning}) {
1618 my @ret_cols = @{$opts->{returning}};
1620 my @ret_vals = try {
1621 local $SIG{__WARN__} = sub {};
1622 my @r = $sth->fetchrow_array;
1628 @ret{@ret_cols} = @ret_vals if (@ret_vals);
1636 return $updated_cols;
1639 ## Currently it is assumed that all values passed will be "normal", i.e. not
1640 ## scalar refs, or at least, all the same type as the first set, the statement is
1641 ## only prepped once.
1643 my ($self, $source, $cols, $data) = @_;
1646 @colvalues{@$cols} = (0..$#$cols);
1648 for my $i (0..$#$cols) {
1649 my $first_val = $data->[0][$i];
1650 next unless ref $first_val eq 'SCALAR';
1652 $colvalues{ $cols->[$i] } = $first_val;
1655 # check for bad data and stringify stringifiable objects
1656 my $bad_slice = sub {
1657 my ($msg, $col_idx, $slice_idx) = @_;
1658 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
1662 local $Data::Dumper::Maxdepth = 1; # don't dump objects, if any
1664 map { $cols->[$_] => $data->[$slice_idx][$_] } (0 .. $#$cols)
1670 for my $datum_idx (0..$#$data) {
1671 my $datum = $data->[$datum_idx];
1673 for my $col_idx (0..$#$cols) {
1674 my $val = $datum->[$col_idx];
1675 my $sqla_bind = $colvalues{ $cols->[$col_idx] };
1676 my $is_literal_sql = (ref $sqla_bind) eq 'SCALAR';
1678 if ($is_literal_sql) {
1680 $bad_slice->('bind found where literal SQL expected', $col_idx, $datum_idx);
1682 elsif ((my $reftype = ref $val) ne 'SCALAR') {
1683 $bad_slice->("$reftype reference found where literal SQL expected",
1684 $col_idx, $datum_idx);
1686 elsif ($$val ne $$sqla_bind){
1687 $bad_slice->("inconsistent literal SQL value, expecting: '$$sqla_bind'",
1688 $col_idx, $datum_idx);
1691 elsif (my $reftype = ref $val) {
1693 if (overload::Method($val, '""')) {
1694 $datum->[$col_idx] = "".$val;
1697 $bad_slice->("$reftype reference found where bind expected",
1698 $col_idx, $datum_idx);
1704 my ($sql, $bind) = $self->_prep_for_execute (
1705 'insert', undef, $source, [\%colvalues]
1709 my $empty_bind = 1 if (not @bind) &&
1710 (grep { ref $_ eq 'SCALAR' } values %colvalues) == @$cols;
1712 if ((not @bind) && (not $empty_bind)) {
1713 $self->throw_exception(
1714 'Cannot insert_bulk without support for placeholders'
1718 # neither _execute_array, nor _execute_inserts_with_no_binds are
1719 # atomic (even if _execute _array is a single call). Thus a safety
1721 my $guard = $self->txn_scope_guard;
1723 $self->_query_start( $sql, [ dummy => '__BULK_INSERT__' ] );
1724 my $sth = $self->sth($sql);
1727 # bind_param_array doesn't work if there are no binds
1728 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
1731 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1732 $self->_execute_array( $source, $sth, \@bind, $cols, $data );
1736 $self->_query_end( $sql, [ dummy => '__BULK_INSERT__' ] );
1740 return (wantarray ? ($rv, $sth, @bind) : $rv);
1743 sub _execute_array {
1744 my ($self, $source, $sth, $bind, $cols, $data, @extra) = @_;
1746 ## This must be an arrayref, else nothing works!
1747 my $tuple_status = [];
1749 ## Get the bind_attributes, if any exist
1750 my $bind_attributes = $self->source_bind_attributes($source);
1752 ## Bind the values and execute
1753 my $placeholder_index = 1;
1755 foreach my $bound (@$bind) {
1757 my $attributes = {};
1758 my ($column_name, $data_index) = @$bound;
1760 if( $bind_attributes ) {
1761 $attributes = $bind_attributes->{$column_name}
1762 if defined $bind_attributes->{$column_name};
1765 my @data = map { $_->[$data_index] } @$data;
1767 $sth->bind_param_array(
1770 (%$attributes ? $attributes : ()),
1772 $placeholder_index++;
1777 $rv = $self->_dbh_execute_array($sth, $tuple_status, @extra);
1783 # Statement must finish even if there was an exception.
1788 $err = shift unless defined $err
1793 if (! defined $err and $sth->err);
1797 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1799 $self->throw_exception("Unexpected populate error: $err")
1800 if ($i > $#$tuple_status);
1802 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1803 ($tuple_status->[$i][1] || $err),
1804 Dumper { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) },
1811 sub _dbh_execute_array {
1812 my ($self, $sth, $tuple_status, @extra) = @_;
1814 return $sth->execute_array({ArrayTupleStatus => $tuple_status});
1817 sub _dbh_execute_inserts_with_no_binds {
1818 my ($self, $sth, $count) = @_;
1822 my $dbh = $self->_get_dbh;
1823 local $dbh->{RaiseError} = 1;
1824 local $dbh->{PrintError} = 0;
1826 $sth->execute foreach 1..$count;
1832 # Make sure statement is finished even if there was an exception.
1837 $err = shift unless defined $err;
1841 $self->throw_exception($err) if defined $err;
1847 my ($self, $source, @args) = @_;
1849 my $bind_attrs = $self->source_bind_attributes($source);
1851 return $self->_execute('update' => [], $source, $bind_attrs, @args);
1856 my ($self, $source, @args) = @_;
1858 my $bind_attrs = $self->source_bind_attributes($source);
1860 return $self->_execute('delete' => [], $source, $bind_attrs, @args);
1863 # We were sent here because the $rs contains a complex search
1864 # which will require a subquery to select the correct rows
1865 # (i.e. joined or limited resultsets, or non-introspectable conditions)
1867 # Generating a single PK column subquery is trivial and supported
1868 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1869 # Look at _multipk_update_delete()
1870 sub _subq_update_delete {
1872 my ($rs, $op, $values) = @_;
1874 my $rsrc = $rs->result_source;
1876 # quick check if we got a sane rs on our hands
1877 my @pcols = $rsrc->_pri_cols;
1879 my $sel = $rs->_resolved_attrs->{select};
1880 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1883 join ("\x00", map { join '.', $rs->{attrs}{alias}, $_ } sort @pcols)
1885 join ("\x00", sort @$sel )
1887 $self->throw_exception (
1888 '_subq_update_delete can not be called on resultsets selecting columns other than the primary keys'
1895 $op eq 'update' ? $values : (),
1896 { $pcols[0] => { -in => $rs->as_query } },
1901 return $self->_multipk_update_delete (@_);
1905 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1906 # resultset update/delete involving subqueries. So by default resort
1907 # to simple (and inefficient) delete_all style per-row opearations,
1908 # while allowing specific storages to override this with a faster
1911 sub _multipk_update_delete {
1912 return shift->_per_row_update_delete (@_);
1915 # This is the default loop used to delete/update rows for multi PK
1916 # resultsets, and used by mysql exclusively (because it can't do anything
1919 # We do not use $row->$op style queries, because resultset update/delete
1920 # is not expected to cascade (this is what delete_all/update_all is for).
1922 # There should be no race conditions as the entire operation is rolled
1925 sub _per_row_update_delete {
1927 my ($rs, $op, $values) = @_;
1929 my $rsrc = $rs->result_source;
1930 my @pcols = $rsrc->_pri_cols;
1932 my $guard = $self->txn_scope_guard;
1934 # emulate the return value of $sth->execute for non-selects
1935 my $row_cnt = '0E0';
1937 my $subrs_cur = $rs->cursor;
1938 my @all_pk = $subrs_cur->all;
1939 for my $pks ( @all_pk) {
1942 for my $i (0.. $#pcols) {
1943 $cond->{$pcols[$i]} = $pks->[$i];
1948 $op eq 'update' ? $values : (),
1962 $self->_execute($self->_select_args(@_));
1965 sub _select_args_to_query {
1968 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $rs_attrs, $rows, $offset)
1969 # = $self->_select_args($ident, $select, $cond, $attrs);
1970 my ($op, $bind, $ident, $bind_attrs, @args) =
1971 $self->_select_args(@_);
1973 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
1974 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1975 $prepared_bind ||= [];
1978 ? ($sql, $prepared_bind, $bind_attrs)
1979 : \[ "($sql)", @$prepared_bind ]
1984 my ($self, $ident, $select, $where, $attrs) = @_;
1986 my $sql_maker = $self->sql_maker;
1987 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1994 $rs_alias && $alias2source->{$rs_alias}
1995 ? ( _rsroot_source_handle => $alias2source->{$rs_alias}->handle )
2000 # calculate bind_attrs before possible $ident mangling
2001 my $bind_attrs = {};
2002 for my $alias (keys %$alias2source) {
2003 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
2004 for my $col (keys %$bindtypes) {
2006 my $fqcn = join ('.', $alias, $col);
2007 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
2009 # Unqialified column names are nice, but at the same time can be
2010 # rather ambiguous. What we do here is basically go along with
2011 # the loop, adding an unqualified column slot to $bind_attrs,
2012 # alongside the fully qualified name. As soon as we encounter
2013 # another column by that name (which would imply another table)
2014 # we unset the unqualified slot and never add any info to it
2015 # to avoid erroneous type binding. If this happens the users
2016 # only choice will be to fully qualify his column name
2018 if (exists $bind_attrs->{$col}) {
2019 $bind_attrs->{$col} = {};
2022 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
2027 # Sanity check the attributes (SQLMaker does it too, but
2028 # in case of a software_limit we'll never reach there)
2029 if (defined $attrs->{offset}) {
2030 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
2031 if ( $attrs->{offset} =~ /\D/ or $attrs->{offset} < 0 );
2033 $attrs->{offset} ||= 0;
2035 if (defined $attrs->{rows}) {
2036 $self->throw_exception("The rows attribute must be a positive integer if present")
2037 if ( $attrs->{rows} =~ /\D/ or $attrs->{rows} <= 0 );
2039 elsif ($attrs->{offset}) {
2040 # MySQL actually recommends this approach. I cringe.
2041 $attrs->{rows} = $sql_maker->__max_int;
2046 # see if we need to tear the prefetch apart otherwise delegate the limiting to the
2047 # storage, unless software limit was requested
2050 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
2052 # grouped prefetch (to satisfy group_by == select)
2053 ( $attrs->{group_by}
2055 @{$attrs->{group_by}}
2057 $attrs->{_prefetch_select}
2059 @{$attrs->{_prefetch_select}}
2062 ($ident, $select, $where, $attrs)
2063 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
2065 elsif (! $attrs->{software_limit} ) {
2066 push @limit, $attrs->{rows}, $attrs->{offset};
2069 # try to simplify the joinmap further (prune unreferenced type-single joins)
2070 $ident = $self->_prune_unused_joins ($ident, $select, $where, $attrs);
2073 # This would be the point to deflate anything found in $where
2074 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2075 # expect a row object. And all we have is a resultsource (it is trivial
2076 # to extract deflator coderefs via $alias2source above).
2078 # I don't see a way forward other than changing the way deflators are
2079 # invoked, and that's just bad...
2082 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $attrs, @limit);
2085 # Returns a counting SELECT for a simple count
2086 # query. Abstracted so that a storage could override
2087 # this to { count => 'firstcol' } or whatever makes
2088 # sense as a performance optimization
2090 #my ($self, $source, $rs_attrs) = @_;
2091 return { count => '*' };
2095 sub source_bind_attributes {
2096 my ($self, $source) = @_;
2098 my $bind_attributes;
2099 foreach my $column ($source->columns) {
2101 my $data_type = $source->column_info($column)->{data_type} || '';
2102 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
2106 return $bind_attributes;
2113 =item Arguments: $ident, $select, $condition, $attrs
2117 Handle a SQL select statement.
2123 my ($ident, $select, $condition, $attrs) = @_;
2124 return $self->cursor_class->new($self, \@_, $attrs);
2129 my ($rv, $sth, @bind) = $self->_select(@_);
2130 my @row = $sth->fetchrow_array;
2131 my @nextrow = $sth->fetchrow_array if @row;
2132 if(@row && @nextrow) {
2133 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2135 # Need to call finish() to work round broken DBDs
2140 =head2 sql_limit_dialect
2142 This is an accessor for the default SQL limit dialect used by a particular
2143 storage driver. Can be overriden by supplying an explicit L</limit_dialect>
2144 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2145 see L<DBIx::Class::SQLMaker::LimitDialects>.
2151 =item Arguments: $sql
2155 Returns a L<DBI> sth (statement handle) for the supplied SQL.
2160 my ($self, $dbh, $sql) = @_;
2162 # 3 is the if_active parameter which avoids active sth re-use
2163 my $sth = $self->disable_sth_caching
2164 ? $dbh->prepare($sql)
2165 : $dbh->prepare_cached($sql, {}, 3);
2167 # XXX You would think RaiseError would make this impossible,
2168 # but apparently that's not true :(
2169 $self->throw_exception($dbh->errstr) if !$sth;
2175 my ($self, $sql) = @_;
2176 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2179 sub _dbh_columns_info_for {
2180 my ($self, $dbh, $table) = @_;
2182 if ($dbh->can('column_info')) {
2186 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2187 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2189 while ( my $info = $sth->fetchrow_hashref() ){
2191 $column_info{data_type} = $info->{TYPE_NAME};
2192 $column_info{size} = $info->{COLUMN_SIZE};
2193 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2194 $column_info{default_value} = $info->{COLUMN_DEF};
2195 my $col_name = $info->{COLUMN_NAME};
2196 $col_name =~ s/^\"(.*)\"$/$1/;
2198 $result{$col_name} = \%column_info;
2203 return \%result if !$caught && scalar keys %result;
2207 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2209 my @columns = @{$sth->{NAME_lc}};
2210 for my $i ( 0 .. $#columns ){
2212 $column_info{data_type} = $sth->{TYPE}->[$i];
2213 $column_info{size} = $sth->{PRECISION}->[$i];
2214 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2216 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2217 $column_info{data_type} = $1;
2218 $column_info{size} = $2;
2221 $result{$columns[$i]} = \%column_info;
2225 foreach my $col (keys %result) {
2226 my $colinfo = $result{$col};
2227 my $type_num = $colinfo->{data_type};
2229 if(defined $type_num && $dbh->can('type_info')) {
2230 my $type_info = $dbh->type_info($type_num);
2231 $type_name = $type_info->{TYPE_NAME} if $type_info;
2232 $colinfo->{data_type} = $type_name if $type_name;
2239 sub columns_info_for {
2240 my ($self, $table) = @_;
2241 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2244 =head2 last_insert_id
2246 Return the row id of the last insert.
2250 sub _dbh_last_insert_id {
2251 my ($self, $dbh, $source, $col) = @_;
2253 my $id = try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2255 return $id if defined $id;
2257 my $class = ref $self;
2258 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2261 sub last_insert_id {
2263 $self->_dbh_last_insert_id ($self->_dbh, @_);
2266 =head2 _native_data_type
2270 =item Arguments: $type_name
2274 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2275 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2276 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2278 The default implementation returns C<undef>, implement in your Storage driver if
2279 you need this functionality.
2281 Should map types from other databases to the native RDBMS type, for example
2282 C<VARCHAR2> to C<VARCHAR>.
2284 Types with modifiers should map to the underlying data type. For example,
2285 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2287 Composite types should map to the container type, for example
2288 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2292 sub _native_data_type {
2293 #my ($self, $data_type) = @_;
2297 # Check if placeholders are supported at all
2298 sub _determine_supports_placeholders {
2300 my $dbh = $self->_get_dbh;
2302 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2303 # but it is inaccurate more often than not
2305 local $dbh->{PrintError} = 0;
2306 local $dbh->{RaiseError} = 1;
2307 $dbh->do('select ?', {}, 1);
2315 # Check if placeholders bound to non-string types throw exceptions
2317 sub _determine_supports_typeless_placeholders {
2319 my $dbh = $self->_get_dbh;
2322 local $dbh->{PrintError} = 0;
2323 local $dbh->{RaiseError} = 1;
2324 # this specifically tests a bind that is NOT a string
2325 $dbh->do('select 1 where 1 = ?', {}, 1);
2335 Returns the database driver name.
2340 shift->_get_dbh->{Driver}->{Name};
2343 =head2 bind_attribute_by_data_type
2345 Given a datatype from column info, returns a database specific bind
2346 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2347 let the database planner just handle it.
2349 Generally only needed for special case column types, like bytea in postgres.
2353 sub bind_attribute_by_data_type {
2357 =head2 is_datatype_numeric
2359 Given a datatype from column_info, returns a boolean value indicating if
2360 the current RDBMS considers it a numeric value. This controls how
2361 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2362 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2363 be performed instead of the usual C<eq>.
2367 sub is_datatype_numeric {
2368 my ($self, $dt) = @_;
2370 return 0 unless $dt;
2372 return $dt =~ /^ (?:
2373 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2378 =head2 create_ddl_dir
2382 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2386 Creates a SQL file based on the Schema, for each of the specified
2387 database engines in C<\@databases> in the given directory.
2388 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2390 Given a previous version number, this will also create a file containing
2391 the ALTER TABLE statements to transform the previous schema into the
2392 current one. Note that these statements may contain C<DROP TABLE> or
2393 C<DROP COLUMN> statements that can potentially destroy data.
2395 The file names are created using the C<ddl_filename> method below, please
2396 override this method in your schema if you would like a different file
2397 name format. For the ALTER file, the same format is used, replacing
2398 $version in the name with "$preversion-$version".
2400 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2401 The most common value for this would be C<< { add_drop_table => 1 } >>
2402 to have the SQL produced include a C<DROP TABLE> statement for each table
2403 created. For quoting purposes supply C<quote_table_names> and
2404 C<quote_field_names>.
2406 If no arguments are passed, then the following default values are assumed:
2410 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2412 =item version - $schema->schema_version
2414 =item directory - './'
2416 =item preversion - <none>
2420 By default, C<\%sqlt_args> will have
2422 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2424 merged with the hash passed in. To disable any of those features, pass in a
2425 hashref like the following
2427 { ignore_constraint_names => 0, # ... other options }
2430 WARNING: You are strongly advised to check all SQL files created, before applying
2435 sub create_ddl_dir {
2436 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2439 carp "No directory given, using ./\n";
2444 make_path ("$dir") # make_path does not like objects (i.e. Path::Class::Dir)
2446 $self->throw_exception(
2447 "Failed to create '$dir': " . ($! || $@ || 'error unknow')
2451 $self->throw_exception ("Directory '$dir' does not exist\n") unless(-d $dir);
2453 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2454 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2456 my $schema_version = $schema->schema_version || '1.x';
2457 $version ||= $schema_version;
2460 add_drop_table => 1,
2461 ignore_constraint_names => 1,
2462 ignore_index_names => 1,
2466 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy')) {
2467 $self->throw_exception("Can't create a ddl file without " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2470 my $sqlt = SQL::Translator->new( $sqltargs );
2472 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2473 my $sqlt_schema = $sqlt->translate({ data => $schema })
2474 or $self->throw_exception ($sqlt->error);
2476 foreach my $db (@$databases) {
2478 $sqlt->{schema} = $sqlt_schema;
2479 $sqlt->producer($db);
2482 my $filename = $schema->ddl_filename($db, $version, $dir);
2483 if (-e $filename && ($version eq $schema_version )) {
2484 # if we are dumping the current version, overwrite the DDL
2485 carp "Overwriting existing DDL file - $filename";
2489 my $output = $sqlt->translate;
2491 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2494 if(!open($file, ">$filename")) {
2495 $self->throw_exception("Can't open $filename for writing ($!)");
2498 print $file $output;
2501 next unless ($preversion);
2503 require SQL::Translator::Diff;
2505 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2506 if(!-e $prefilename) {
2507 carp("No previous schema file found ($prefilename)");
2511 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2513 carp("Overwriting existing diff file - $difffile");
2519 my $t = SQL::Translator->new($sqltargs);
2524 or $self->throw_exception ($t->error);
2526 my $out = $t->translate( $prefilename )
2527 or $self->throw_exception ($t->error);
2529 $source_schema = $t->schema;
2531 $source_schema->name( $prefilename )
2532 unless ( $source_schema->name );
2535 # The "new" style of producers have sane normalization and can support
2536 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2537 # And we have to diff parsed SQL against parsed SQL.
2538 my $dest_schema = $sqlt_schema;
2540 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2541 my $t = SQL::Translator->new($sqltargs);
2546 or $self->throw_exception ($t->error);
2548 my $out = $t->translate( $filename )
2549 or $self->throw_exception ($t->error);
2551 $dest_schema = $t->schema;
2553 $dest_schema->name( $filename )
2554 unless $dest_schema->name;
2557 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2561 if(!open $file, ">$difffile") {
2562 $self->throw_exception("Can't write to $difffile ($!)");
2570 =head2 deployment_statements
2574 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2578 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2580 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2581 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2583 C<$directory> is used to return statements from files in a previously created
2584 L</create_ddl_dir> directory and is optional. The filenames are constructed
2585 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2587 If no C<$directory> is specified then the statements are constructed on the
2588 fly using L<SQL::Translator> and C<$version> is ignored.
2590 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2594 sub deployment_statements {
2595 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2596 $type ||= $self->sqlt_type;
2597 $version ||= $schema->schema_version || '1.x';
2599 my $filename = $schema->ddl_filename($type, $version, $dir);
2603 open($file, "<$filename")
2604 or $self->throw_exception("Can't open $filename ($!)");
2607 return join('', @rows);
2610 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy') ) {
2611 $self->throw_exception("Can't deploy without a ddl_dir or " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2614 # sources needs to be a parser arg, but for simplicty allow at top level
2616 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2617 if exists $sqltargs->{sources};
2619 my $tr = SQL::Translator->new(
2620 producer => "SQL::Translator::Producer::${type}",
2622 parser => 'SQL::Translator::Parser::DBIx::Class',
2629 @ret = $tr->translate;
2632 $ret[0] = $tr->translate;
2635 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
2636 unless (@ret && defined $ret[0]);
2638 return $wa ? @ret : $ret[0];
2642 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2645 return if($line =~ /^--/);
2647 # next if($line =~ /^DROP/m);
2648 return if($line =~ /^BEGIN TRANSACTION/m);
2649 return if($line =~ /^COMMIT/m);
2650 return if $line =~ /^\s+$/; # skip whitespace only
2651 $self->_query_start($line);
2653 # do a dbh_do cycle here, as we need some error checking in
2654 # place (even though we will ignore errors)
2655 $self->dbh_do (sub { $_[1]->do($line) });
2657 carp qq{$_ (running "${line}")};
2659 $self->_query_end($line);
2661 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2662 if (@statements > 1) {
2663 foreach my $statement (@statements) {
2664 $deploy->( $statement );
2667 elsif (@statements == 1) {
2668 foreach my $line ( split(";\n", $statements[0])) {
2674 =head2 datetime_parser
2676 Returns the datetime parser class
2680 sub datetime_parser {
2682 return $self->{datetime_parser} ||= do {
2683 $self->build_datetime_parser(@_);
2687 =head2 datetime_parser_type
2689 Defines (returns) the datetime parser class - currently hardwired to
2690 L<DateTime::Format::MySQL>
2694 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2696 =head2 build_datetime_parser
2698 See L</datetime_parser>
2702 sub build_datetime_parser {
2704 my $type = $self->datetime_parser_type(@_);
2705 $self->ensure_class_loaded ($type);
2710 =head2 is_replicating
2712 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2713 replicate from a master database. Default is undef, which is the result
2714 returned by databases that don't support replication.
2718 sub is_replicating {
2723 =head2 lag_behind_master
2725 Returns a number that represents a certain amount of lag behind a master db
2726 when a given storage is replicating. The number is database dependent, but
2727 starts at zero and increases with the amount of lag. Default in undef
2731 sub lag_behind_master {
2735 =head2 relname_to_table_alias
2739 =item Arguments: $relname, $join_count
2743 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
2746 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
2747 way these aliases are named.
2749 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
2750 otherwise C<"$relname">.
2754 sub relname_to_table_alias {
2755 my ($self, $relname, $join_count) = @_;
2757 my $alias = ($join_count && $join_count > 1 ?
2758 join('_', $relname, $join_count) : $relname);
2767 =head2 DBIx::Class and AutoCommit
2769 DBIx::Class can do some wonderful magic with handling exceptions,
2770 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2771 (the default) combined with C<txn_do> for transaction support.
2773 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2774 in an assumed transaction between commits, and you're telling us you'd
2775 like to manage that manually. A lot of the magic protections offered by
2776 this module will go away. We can't protect you from exceptions due to database
2777 disconnects because we don't know anything about how to restart your
2778 transactions. You're on your own for handling all sorts of exceptional
2779 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2785 Matt S. Trout <mst@shadowcatsystems.co.uk>
2787 Andy Grundman <andy@hybridized.org>
2791 You may distribute this code under the same terms as Perl itself.