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 DBIx::Class::Carp;
11 use Scalar::Util qw/refaddr weaken reftype blessed/;
12 use List::Util qw/first/;
13 use Sub::Name 'subname';
14 use Context::Preserve 'preserve_context';
17 use Data::Compare (); # no imports!!! guard against insane architecture
20 # default cursor class, overridable in connect_info attributes
21 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
23 __PACKAGE__->mk_group_accessors('inherited' => qw/
24 sql_limit_dialect sql_quote_char sql_name_sep
27 __PACKAGE__->mk_group_accessors('component_class' => qw/sql_maker_class datetime_parser_type/);
29 __PACKAGE__->sql_maker_class('DBIx::Class::SQLMaker');
30 __PACKAGE__->datetime_parser_type('DateTime::Format::MySQL'); # historic default
32 __PACKAGE__->sql_name_sep('.');
34 __PACKAGE__->mk_group_accessors('simple' => qw/
35 _connect_info _dbic_connect_attributes _driver_determined
36 _dbh _dbh_details _conn_pid _sql_maker _sql_maker_opts _dbh_autocommit
37 _perform_autoinc_retrieval _autoinc_supplied_for_op
40 # the values for these accessors are picked out (and deleted) from
41 # the attribute hashref passed to connect_info
42 my @storage_options = qw/
43 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
44 disable_sth_caching unsafe auto_savepoint
46 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
49 # capability definitions, using a 2-tiered accessor system
52 # A driver/user may define _use_X, which blindly without any checks says:
53 # "(do not) use this capability", (use_dbms_capability is an "inherited"
56 # If _use_X is undef, _supports_X is then queried. This is a "simple" style
57 # accessor, which in turn calls _determine_supports_X, and stores the return
58 # in a special slot on the storage object, which is wiped every time a $dbh
59 # reconnection takes place (it is not guaranteed that upon reconnection we
60 # will get the same rdbms version). _determine_supports_X does not need to
61 # exist on a driver, as we ->can for it before calling.
63 my @capabilities = (qw/
65 insert_returning_bound
74 __PACKAGE__->mk_group_accessors( dbms_capability => map { "_supports_$_" } @capabilities );
75 __PACKAGE__->mk_group_accessors( use_dbms_capability => map { "_use_$_" } (@capabilities ) );
77 # on by default, not strictly a capability (pending rewrite)
78 __PACKAGE__->_use_join_optimizer (1);
79 sub _determine_supports_join_optimizer { 1 };
81 # Each of these methods need _determine_driver called before itself
82 # in order to function reliably. This is a purely DRY optimization
84 # get_(use)_dbms_capability need to be called on the correct Storage
85 # class, as _use_X may be hardcoded class-wide, and _supports_X calls
86 # _determine_supports_X which obv. needs a correct driver as well
87 my @rdbms_specific_methods = qw/
106 with_deferred_fk_checks
108 get_use_dbms_capability
115 for my $meth (@rdbms_specific_methods) {
117 my $orig = __PACKAGE__->can ($meth)
118 or die "$meth is not a ::Storage::DBI method!";
121 no warnings qw/redefine/;
122 *{__PACKAGE__ ."::$meth"} = subname $meth => sub {
124 # only fire when invoked on an instance, a valid class-based invocation
125 # would e.g. be setting a default for an inherited accessor
128 ! $_[0]->{_driver_determined}
130 ! $_[0]->{_in_determine_driver}
132 # Only try to determine stuff if we have *something* that either is or can
133 # provide a DSN. Allows for bare $schema's generated with a plain ->connect()
134 # to still be marginally useful
135 $_[0]->_dbi_connect_info->[0]
137 $_[0]->_determine_driver;
139 # This for some reason crashes and burns on perl 5.8.1
140 # IFF the method ends up throwing an exception
141 #goto $_[0]->can ($meth);
143 my $cref = $_[0]->can ($meth);
153 DBIx::Class::Storage::DBI - DBI storage handler
157 my $schema = MySchema->connect('dbi:SQLite:my.db');
159 $schema->storage->debug(1);
161 my @stuff = $schema->storage->dbh_do(
163 my ($storage, $dbh, @args) = @_;
164 $dbh->do("DROP TABLE authors");
169 $schema->resultset('Book')->search({
170 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
175 This class represents the connection to an RDBMS via L<DBI>. See
176 L<DBIx::Class::Storage> for general information. This pod only
177 documents DBI-specific methods and behaviors.
184 my $new = shift->next::method(@_);
186 $new->_sql_maker_opts({});
187 $new->_dbh_details({});
188 $new->{_in_do_block} = 0;
190 # read below to see what this does
191 $new->_arm_global_destructor;
196 # This is hack to work around perl shooting stuff in random
197 # order on exit(). If we do not walk the remaining storage
198 # objects in an END block, there is a *small but real* chance
199 # of a fork()ed child to kill the parent's shared DBI handle,
200 # *before perl reaches the DESTROY in this package*
201 # Yes, it is ugly and effective.
202 # Additionally this registry is used by the CLONE method to
203 # make sure no handles are shared between threads
205 my %seek_and_destroy;
207 sub _arm_global_destructor {
209 # quick "garbage collection" pass - prevents the registry
210 # from slowly growing with a bunch of undef-valued keys
211 defined $seek_and_destroy{$_} or delete $seek_and_destroy{$_}
212 for keys %seek_and_destroy;
215 $seek_and_destroy{ refaddr($_[0]) } = $_[0]
220 local $?; # just in case the DBI destructor changes it somehow
222 # destroy just the object if not native to this process
223 $_->_verify_pid for (grep
225 values %seek_and_destroy
230 # As per DBI's recommendation, DBIC disconnects all handles as
231 # soon as possible (DBIC will reconnect only on demand from within
233 my @instances = grep { defined $_ } values %seek_and_destroy;
234 %seek_and_destroy = ();
239 $_->transaction_depth(0);
242 # properly renumber existing refs
243 $_->_arm_global_destructor
251 # some databases spew warnings on implicit disconnect
252 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
253 local $SIG{__WARN__} = sub {};
256 # this op is necessary, since the very last perl runtime statement
257 # triggers a global destruction shootout, and the $SIG localization
258 # may very well be destroyed before perl actually gets to do the
263 # handle pid changes correctly - do not destroy parent's connection
267 my $pid = $self->_conn_pid;
268 if( defined $pid and $pid != $$ and my $dbh = $self->_dbh ) {
269 $dbh->{InactiveDestroy} = 1;
271 $self->transaction_depth(0);
272 $self->savepoints([]);
280 This method is normally called by L<DBIx::Class::Schema/connection>, which
281 encapsulates its argument list in an arrayref before passing them here.
283 The argument list may contain:
289 The same 4-element argument set one would normally pass to
290 L<DBI/connect>, optionally followed by
291 L<extra attributes|/DBIx::Class specific connection attributes>
292 recognized by DBIx::Class:
294 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
298 A single code reference which returns a connected
299 L<DBI database handle|DBI/connect> optionally followed by
300 L<extra attributes|/DBIx::Class specific connection attributes> recognized
303 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
307 A single hashref with all the attributes and the dsn/user/password
310 $connect_info_args = [{
318 $connect_info_args = [{
319 dbh_maker => sub { DBI->connect (...) },
324 This is particularly useful for L<Catalyst> based applications, allowing the
325 following config (L<Config::General> style):
330 dsn dbi:mysql:database=test
337 The C<dsn>/C<user>/C<password> combination can be substituted by the
338 C<dbh_maker> key whose value is a coderef that returns a connected
339 L<DBI database handle|DBI/connect>
343 Please note that the L<DBI> docs recommend that you always explicitly
344 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
345 recommends that it be set to I<1>, and that you perform transactions
346 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
347 to I<1> if you do not do explicitly set it to zero. This is the default
348 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
350 =head3 DBIx::Class specific connection attributes
352 In addition to the standard L<DBI|DBI/ATTRIBUTES COMMON TO ALL HANDLES>
353 L<connection|DBI/Database Handle Attributes> attributes, DBIx::Class recognizes
354 the following connection options. These options can be mixed in with your other
355 L<DBI> connection attributes, or placed in a separate hashref
356 (C<\%extra_attributes>) as shown above.
358 Every time C<connect_info> is invoked, any previous settings for
359 these options will be cleared before setting the new ones, regardless of
360 whether any options are specified in the new C<connect_info>.
367 Specifies things to do immediately after connecting or re-connecting to
368 the database. Its value may contain:
374 This contains one SQL statement to execute.
376 =item an array reference
378 This contains SQL statements to execute in order. Each element contains
379 a string or a code reference that returns a string.
381 =item a code reference
383 This contains some code to execute. Unlike code references within an
384 array reference, its return value is ignored.
388 =item on_disconnect_do
390 Takes arguments in the same form as L</on_connect_do> and executes them
391 immediately before disconnecting from the database.
393 Note, this only runs if you explicitly call L</disconnect> on the
396 =item on_connect_call
398 A more generalized form of L</on_connect_do> that calls the specified
399 C<connect_call_METHOD> methods in your storage driver.
401 on_connect_do => 'select 1'
405 on_connect_call => [ [ do_sql => 'select 1' ] ]
407 Its values may contain:
413 Will call the C<connect_call_METHOD> method.
415 =item a code reference
417 Will execute C<< $code->($storage) >>
419 =item an array reference
421 Each value can be a method name or code reference.
423 =item an array of arrays
425 For each array, the first item is taken to be the C<connect_call_> method name
426 or code reference, and the rest are parameters to it.
430 Some predefined storage methods you may use:
436 Executes a SQL string or a code reference that returns a SQL string. This is
437 what L</on_connect_do> and L</on_disconnect_do> use.
445 Will execute the scalar as SQL.
449 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
450 attributes hashref and bind values.
452 =item a code reference
454 Will execute C<< $code->($storage) >> and execute the return array refs as
461 Execute any statements necessary to initialize the database session to return
462 and accept datetime/timestamp values used with
463 L<DBIx::Class::InflateColumn::DateTime>.
465 Only necessary for some databases, see your specific storage driver for
466 implementation details.
470 =item on_disconnect_call
472 Takes arguments in the same form as L</on_connect_call> and executes them
473 immediately before disconnecting from the database.
475 Calls the C<disconnect_call_METHOD> methods as opposed to the
476 C<connect_call_METHOD> methods called by L</on_connect_call>.
478 Note, this only runs if you explicitly call L</disconnect> on the
481 =item disable_sth_caching
483 If set to a true value, this option will disable the caching of
484 statement handles via L<DBI/prepare_cached>.
488 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
489 default L</sql_limit_dialect> setting of the storage (if any). For a list
490 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
494 When true automatically sets L</quote_char> and L</name_sep> to the characters
495 appropriate for your particular RDBMS. This option is preferred over specifying
496 L</quote_char> directly.
500 Specifies what characters to use to quote table and column names.
502 C<quote_char> expects either a single character, in which case is it
503 is placed on either side of the table/column name, or an arrayref of length
504 2 in which case the table/column name is placed between the elements.
506 For example under MySQL you should use C<< quote_char => '`' >>, and for
507 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
511 This parameter is only useful in conjunction with C<quote_char>, and is used to
512 specify the character that separates elements (schemas, tables, columns) from
513 each other. If unspecified it defaults to the most commonly used C<.>.
517 This Storage driver normally installs its own C<HandleError>, sets
518 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
519 all database handles, including those supplied by a coderef. It does this
520 so that it can have consistent and useful error behavior.
522 If you set this option to a true value, Storage will not do its usual
523 modifications to the database handle's attributes, and instead relies on
524 the settings in your connect_info DBI options (or the values you set in
525 your connection coderef, in the case that you are connecting via coderef).
527 Note that your custom settings can cause Storage to malfunction,
528 especially if you set a C<HandleError> handler that suppresses exceptions
529 and/or disable C<RaiseError>.
533 If this option is true, L<DBIx::Class> will use savepoints when nesting
534 transactions, making it possible to recover from failure in the inner
535 transaction without having to abort all outer transactions.
539 Use this argument to supply a cursor class other than the default
540 L<DBIx::Class::Storage::DBI::Cursor>.
544 Some real-life examples of arguments to L</connect_info> and
545 L<DBIx::Class::Schema/connect>
547 # Simple SQLite connection
548 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
551 ->connect_info([ sub { DBI->connect(...) } ]);
553 # Connect via subref in hashref
555 dbh_maker => sub { DBI->connect(...) },
556 on_connect_do => 'alter session ...',
559 # A bit more complicated
566 { quote_char => q{"} },
570 # Equivalent to the previous example
576 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
580 # Same, but with hashref as argument
581 # See parse_connect_info for explanation
584 dsn => 'dbi:Pg:dbname=foo',
586 password => 'my_pg_password',
593 # Subref + DBIx::Class-specific connection options
596 sub { DBI->connect(...) },
600 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
601 disable_sth_caching => 1,
611 my ($self, $info) = @_;
613 return $self->_connect_info if !$info;
615 $self->_connect_info($info); # copy for _connect_info
617 $info = $self->_normalize_connect_info($info)
618 if ref $info eq 'ARRAY';
621 %{ $self->_default_dbi_connect_attributes || {} },
622 %{ $info->{attributes} || {} },
625 my @args = @{ $info->{arguments} };
627 if (keys %attrs and ref $args[0] ne 'CODE') {
629 'You provided explicit AutoCommit => 0 in your connection_info. '
630 . 'This is almost universally a bad idea (see the footnotes of '
631 . 'DBIx::Class::Storage::DBI for more info). If you still want to '
632 . 'do this you can set $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK} to disable '
634 ) if ! $attrs{AutoCommit} and ! $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK};
636 push @args, \%attrs if keys %attrs;
639 # this is the authoritative "always an arrayref" thing fed to DBI->connect
640 # OR a single-element coderef-based $dbh factory
641 $self->_dbi_connect_info(\@args);
643 # extract the individual storage options
644 for my $storage_opt (keys %{ $info->{storage_options} }) {
645 my $value = $info->{storage_options}{$storage_opt};
647 $self->$storage_opt($value);
650 # Extract the individual sqlmaker options
652 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
653 # the new set of options
654 $self->_sql_maker(undef);
655 $self->_sql_maker_opts({});
657 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
658 my $value = $info->{sql_maker_options}{$sql_maker_opt};
660 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
664 # save attributes in a separate accessor so they are always
665 # introspectable, even in case of a CODE $dbhmaker
666 $self->_dbic_connect_attributes (\%attrs);
668 return $self->_connect_info;
671 sub _dbi_connect_info {
674 return $self->{_dbi_connect_info} = $_[0]
677 my $conninfo = $self->{_dbi_connect_info} || [];
679 # last ditch effort to grab a DSN
680 if ( ! defined $conninfo->[0] and $ENV{DBI_DSN} ) {
681 my @new_conninfo = @$conninfo;
682 $new_conninfo[0] = $ENV{DBI_DSN};
683 $conninfo = \@new_conninfo;
690 sub _normalize_connect_info {
691 my ($self, $info_arg) = @_;
694 my @args = @$info_arg; # take a shallow copy for further mutilation
696 # combine/pre-parse arguments depending on invocation style
699 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
700 %attrs = %{ $args[1] || {} };
703 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
704 %attrs = %{$args[0]};
706 if (my $code = delete $attrs{dbh_maker}) {
709 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
712 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
713 . "to the result of 'dbh_maker'",
715 join (', ', map { "'$_'" } (@ignored) ),
720 @args = delete @attrs{qw/dsn user password/};
723 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
725 % { $args[3] || {} },
726 % { $args[4] || {} },
728 @args = @args[0,1,2];
731 $info{arguments} = \@args;
733 my @storage_opts = grep exists $attrs{$_},
734 @storage_options, 'cursor_class';
736 @{ $info{storage_options} }{@storage_opts} =
737 delete @attrs{@storage_opts} if @storage_opts;
739 my @sql_maker_opts = grep exists $attrs{$_},
740 qw/limit_dialect quote_char name_sep quote_names/;
742 @{ $info{sql_maker_options} }{@sql_maker_opts} =
743 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
745 $info{attributes} = \%attrs if %attrs;
750 sub _default_dbi_connect_attributes () {
755 ShowErrorStatement => 1,
761 This method is deprecated in favour of setting via L</connect_info>.
765 =head2 on_disconnect_do
767 This method is deprecated in favour of setting via L</connect_info>.
771 sub _parse_connect_do {
772 my ($self, $type) = @_;
774 my $val = $self->$type;
775 return () if not defined $val;
780 push @res, [ 'do_sql', $val ];
781 } elsif (ref($val) eq 'CODE') {
783 } elsif (ref($val) eq 'ARRAY') {
784 push @res, map { [ 'do_sql', $_ ] } @$val;
786 $self->throw_exception("Invalid type for $type: ".ref($val));
794 Arguments: ($subref | $method_name), @extra_coderef_args?
796 Execute the given $subref or $method_name using the new exception-based
797 connection management.
799 The first two arguments will be the storage object that C<dbh_do> was called
800 on and a database handle to use. Any additional arguments will be passed
801 verbatim to the called subref as arguments 2 and onwards.
803 Using this (instead of $self->_dbh or $self->dbh) ensures correct
804 exception handling and reconnection (or failover in future subclasses).
806 Your subref should have no side-effects outside of the database, as
807 there is the potential for your subref to be partially double-executed
808 if the database connection was stale/dysfunctional.
812 my @stuff = $schema->storage->dbh_do(
814 my ($storage, $dbh, @cols) = @_;
815 my $cols = join(q{, }, @cols);
816 $dbh->selectrow_array("SELECT $cols FROM foo");
825 my $run_target = shift;
827 # short circuit when we know there is no need for a runner
829 # FIXME - assumption may be wrong
830 # the rationale for the txn_depth check is that if this block is a part
831 # of a larger transaction, everything up to that point is screwed anyway
832 return $self->$run_target($self->_get_dbh, @_)
833 if $self->{_in_do_block} or $self->transaction_depth;
835 # take a ref instead of a copy, to preserve @_ aliasing
836 # semantics within the coderef, but only if needed
837 # (pseudoforking doesn't like this trick much)
838 my $args = @_ ? \@_ : [];
840 DBIx::Class::Storage::BlockRunner->new(
842 run_code => sub { $self->$run_target ($self->_get_dbh, @$args ) },
844 retry_handler => sub { ! ( $_[0]->retried_count or $_[0]->storage->connected ) },
849 $_[0]->_get_dbh; # connects or reconnects on pid change, necessary to grab correct txn_depth
850 shift->next::method(@_);
855 Our C<disconnect> method also performs a rollback first if the
856 database is not in C<AutoCommit> mode.
866 push @actions, ( $self->on_disconnect_call || () );
867 push @actions, $self->_parse_connect_do ('on_disconnect_do');
869 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
871 # stops the "implicit rollback on disconnect" warning
872 $self->_exec_txn_rollback unless $self->_dbh_autocommit;
874 %{ $self->_dbh->{CachedKids} } = ();
875 $self->_dbh->disconnect;
880 =head2 with_deferred_fk_checks
884 =item Arguments: C<$coderef>
886 =item Return Value: The return value of $coderef
890 Storage specific method to run the code ref with FK checks deferred or
891 in MySQL's case disabled entirely.
895 # Storage subclasses should override this
896 sub with_deferred_fk_checks {
897 my ($self, $sub) = @_;
905 =item Arguments: none
907 =item Return Value: 1|0
911 Verifies that the current database handle is active and ready to execute
912 an SQL statement (e.g. the connection did not get stale, server is still
913 answering, etc.) This method is used internally by L</dbh>.
919 return 0 unless $self->_seems_connected;
922 local $self->_dbh->{RaiseError} = 1;
927 sub _seems_connected {
930 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
932 my $dbh = $self->_dbh
935 return $dbh->FETCH('Active');
941 my $dbh = $self->_dbh or return 0;
946 sub ensure_connected {
949 unless ($self->connected) {
950 $self->_populate_dbh;
956 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
957 is guaranteed to be healthy by implicitly calling L</connected>, and if
958 necessary performing a reconnection before returning. Keep in mind that this
959 is very B<expensive> on some database engines. Consider using L</dbh_do>
967 if (not $self->_dbh) {
968 $self->_populate_dbh;
970 $self->ensure_connected;
975 # this is the internal "get dbh or connect (don't check)" method
978 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
979 $self->_populate_dbh unless $self->_dbh;
985 unless ($self->_sql_maker) {
986 my $sql_maker_class = $self->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. '
1001 . "Please file an RT ticket against '$s_class'"
1002 ) if $self->_dbi_connect_info->[0];
1008 my ($quote_char, $name_sep);
1010 if ($opts{quote_names}) {
1011 $quote_char = (delete $opts{quote_char}) || $self->sql_quote_char || do {
1012 my $s_class = (ref $self) || $self;
1014 "You requested 'quote_names' but your storage class ($s_class) does "
1015 . 'not explicitly define a default sql_quote_char and you have not '
1016 . 'supplied a quote_char as part of your connection_info. DBIC will '
1017 .q{default to the ANSI SQL standard quote '"', which works most of }
1018 . "the time. Please file an RT ticket against '$s_class'."
1024 $name_sep = (delete $opts{name_sep}) || $self->sql_name_sep;
1027 $self->_sql_maker($sql_maker_class->new(
1028 bindtype=>'columns',
1029 array_datatypes => 1,
1030 limit_dialect => $dialect,
1031 ($quote_char ? (quote_char => $quote_char) : ()),
1032 name_sep => ($name_sep || '.'),
1036 return $self->_sql_maker;
1039 # nothing to do by default
1046 $self->_dbh(undef); # in case ->connected failed we might get sent here
1047 $self->_dbh_details({}); # reset everything we know
1049 $self->_dbh($self->_connect);
1051 $self->_conn_pid($$) unless DBIx::Class::_ENV_::BROKEN_FORK; # on win32 these are in fact threads
1053 $self->_determine_driver;
1055 # Always set the transaction depth on connect, since
1056 # there is no transaction in progress by definition
1057 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1059 $self->_run_connection_actions unless $self->{_in_determine_driver};
1062 sub _run_connection_actions {
1066 push @actions, ( $self->on_connect_call || () );
1067 push @actions, $self->_parse_connect_do ('on_connect_do');
1069 $self->_do_connection_actions(connect_call_ => $_) for @actions;
1074 sub set_use_dbms_capability {
1075 $_[0]->set_inherited ($_[1], $_[2]);
1078 sub get_use_dbms_capability {
1079 my ($self, $capname) = @_;
1081 my $use = $self->get_inherited ($capname);
1084 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1088 sub set_dbms_capability {
1089 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1092 sub get_dbms_capability {
1093 my ($self, $capname) = @_;
1095 my $cap = $self->_dbh_details->{capability}{$capname};
1097 unless (defined $cap) {
1098 if (my $meth = $self->can ("_determine$capname")) {
1099 $cap = $self->$meth ? 1 : 0;
1105 $self->set_dbms_capability ($capname, $cap);
1115 unless ($info = $self->_dbh_details->{info}) {
1119 my $server_version = try {
1120 $self->_get_server_version
1122 # driver determination *may* use this codepath
1123 # in which case we must rethrow
1124 $self->throw_exception($_) if $self->{_in_determine_driver};
1126 # $server_version on failure
1130 if (defined $server_version) {
1131 $info->{dbms_version} = $server_version;
1133 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1134 my @verparts = split (/\./, $numeric_version);
1140 # consider only up to 3 version parts, iff not more than 3 digits
1142 while (@verparts && @use_parts < 3) {
1143 my $p = shift @verparts;
1145 push @use_parts, $p;
1147 push @use_parts, 0 while @use_parts < 3;
1149 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1153 $self->_dbh_details->{info} = $info;
1159 sub _get_server_version {
1160 shift->_dbh_get_info('SQL_DBMS_VER');
1164 my ($self, $info) = @_;
1166 if ($info =~ /[^0-9]/) {
1167 require DBI::Const::GetInfoType;
1168 $info = $DBI::Const::GetInfoType::GetInfoType{$info};
1169 $self->throw_exception("Info type '$_[1]' not provided by DBI::Const::GetInfoType")
1170 unless defined $info;
1173 $self->_get_dbh->get_info($info);
1176 sub _describe_connection {
1177 require DBI::Const::GetInfoReturn;
1183 $drv = $self->_extract_driver_from_connect_info;
1184 $self->ensure_connected;
1187 $drv = "DBD::$drv" if $drv;
1190 DBIC_DSN => $self->_dbi_connect_info->[0],
1191 DBI_VER => DBI->VERSION,
1192 DBIC_VER => DBIx::Class->VERSION,
1193 DBIC_DRIVER => ref $self,
1196 DBD_VER => try { $drv->VERSION },
1200 # try to grab data even if we never managed to connect
1201 # will cover us in cases of an oddly broken half-connect
1203 #keys %DBI::Const::GetInfoType::GetInfoType,
1205 SQL_CURSOR_COMMIT_BEHAVIOR
1206 SQL_CURSOR_ROLLBACK_BEHAVIOR
1207 SQL_CURSOR_SENSITIVITY
1208 SQL_DATA_SOURCE_NAME
1211 SQL_DEFAULT_TXN_ISOLATION
1216 SQL_EXPRESSIONS_IN_ORDERBY
1219 SQL_IDENTIFIER_QUOTE_CHAR
1220 SQL_MAX_CATALOG_NAME_LEN
1221 SQL_MAX_COLUMN_NAME_LEN
1222 SQL_MAX_IDENTIFIER_LEN
1223 SQL_MAX_TABLE_NAME_LEN
1224 SQL_MULTIPLE_ACTIVE_TXN
1225 SQL_MULT_RESULT_SETS
1226 SQL_NEED_LONG_DATA_LEN
1227 SQL_NON_NULLABLE_COLUMNS
1229 SQL_QUALIFIER_NAME_SEPARATOR
1230 SQL_QUOTED_IDENTIFIER_CASE
1232 SQL_TXN_ISOLATION_OPTION
1235 # some drivers barf on things they do not know about instead
1236 # of returning undef
1237 my $v = try { $self->_dbh_get_info($inf) };
1238 next unless defined $v;
1240 #my $key = sprintf( '%s(%s)', $inf, $DBI::Const::GetInfoType::GetInfoType{$inf} );
1241 my $expl = DBI::Const::GetInfoReturn::Explain($inf, $v);
1242 $res->{$inf} = DBI::Const::GetInfoReturn::Format($inf, $v) . ( $expl ? " ($expl)" : '' );
1248 sub _determine_driver {
1251 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1252 my $started_connected = 0;
1253 local $self->{_in_determine_driver} = 1;
1255 if (ref($self) eq __PACKAGE__) {
1257 if ($self->_dbh) { # we are connected
1258 $driver = $self->_dbh->{Driver}{Name};
1259 $started_connected = 1;
1262 $driver = $self->_extract_driver_from_connect_info;
1266 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1267 if ($self->load_optional_class($storage_class)) {
1268 mro::set_mro($storage_class, 'c3');
1269 bless $self, $storage_class;
1273 $self->_warn_undetermined_driver(
1274 'This version of DBIC does not yet seem to supply a driver for '
1275 . "your particular RDBMS and/or connection method ('$driver')."
1280 $self->_warn_undetermined_driver(
1281 'Unable to extract a driver name from connect info - this '
1282 . 'should not have happened.'
1287 $self->_driver_determined(1);
1289 Class::C3->reinitialize() if DBIx::Class::_ENV_::OLD_MRO;
1291 if ($self->can('source_bind_attributes')) {
1292 $self->throw_exception(
1293 "Your storage subclass @{[ ref $self ]} provides (or inherits) the method "
1294 . 'source_bind_attributes() for which support has been removed as of Jan 2013. '
1295 . 'If you are not sure how to proceed please contact the development team via '
1296 . 'http://search.cpan.org/dist/DBIx-Class/lib/DBIx/Class.pm#GETTING_HELP/SUPPORT'
1300 $self->_init; # run driver-specific initializations
1302 $self->_run_connection_actions
1303 if !$started_connected && defined $self->_dbh;
1307 sub _extract_driver_from_connect_info {
1312 # if connect_info is a CODEREF, we have no choice but to connect
1314 ref $self->_dbi_connect_info->[0]
1316 reftype $self->_dbi_connect_info->[0] eq 'CODE'
1318 $self->_populate_dbh;
1319 $drv = $self->_dbh->{Driver}{Name};
1322 # try to use dsn to not require being connected, the driver may still
1323 # force a connection later in _rebless to determine version
1324 # (dsn may not be supplied at all if all we do is make a mock-schema)
1325 ($drv) = ($self->_dbi_connect_info->[0] || '') =~ /^dbi:([^:]+):/i;
1326 $drv ||= $ENV{DBI_DRIVER};
1332 sub _determine_connector_driver {
1333 my ($self, $conn) = @_;
1335 my $dbtype = $self->_dbh_get_info('SQL_DBMS_NAME');
1338 $self->_warn_undetermined_driver(
1339 'Unable to retrieve RDBMS type (SQL_DBMS_NAME) of the engine behind your '
1340 . "$conn connector - this should not have happened."
1345 $dbtype =~ s/\W/_/gi;
1347 my $subclass = "DBIx::Class::Storage::DBI::${conn}::${dbtype}";
1348 return if $self->isa($subclass);
1350 if ($self->load_optional_class($subclass)) {
1351 bless $self, $subclass;
1355 $self->_warn_undetermined_driver(
1356 'This version of DBIC does not yet seem to supply a driver for '
1357 . "your particular RDBMS and/or connection method ('$conn/$dbtype')."
1362 sub _warn_undetermined_driver {
1363 my ($self, $msg) = @_;
1365 require Data::Dumper::Concise;
1367 carp_once ($msg . ' While we will attempt to continue anyway, the results '
1368 . 'are likely to be underwhelming. Please upgrade DBIC, and if this message '
1369 . "does not go away, file a bugreport including the following info:\n"
1370 . Data::Dumper::Concise::Dumper($self->_describe_connection)
1374 sub _do_connection_actions {
1376 my $method_prefix = shift;
1379 if (not ref($call)) {
1380 my $method = $method_prefix . $call;
1382 } elsif (ref($call) eq 'CODE') {
1384 } elsif (ref($call) eq 'ARRAY') {
1385 if (ref($call->[0]) ne 'ARRAY') {
1386 $self->_do_connection_actions($method_prefix, $_) for @$call;
1388 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1391 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1397 sub connect_call_do_sql {
1399 $self->_do_query(@_);
1402 sub disconnect_call_do_sql {
1404 $self->_do_query(@_);
1407 # override in db-specific backend when necessary
1408 sub connect_call_datetime_setup { 1 }
1411 my ($self, $action) = @_;
1413 if (ref $action eq 'CODE') {
1414 $action = $action->($self);
1415 $self->_do_query($_) foreach @$action;
1418 # Most debuggers expect ($sql, @bind), so we need to exclude
1419 # the attribute hash which is the second argument to $dbh->do
1420 # furthermore the bind values are usually to be presented
1421 # as named arrayref pairs, so wrap those here too
1422 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1423 my $sql = shift @do_args;
1424 my $attrs = shift @do_args;
1425 my @bind = map { [ undef, $_ ] } @do_args;
1428 $_[0]->_query_start($sql, \@bind);
1429 $_[1]->do($sql, $attrs, @do_args);
1430 $_[0]->_query_end($sql, \@bind);
1440 my $info = $self->_dbi_connect_info;
1442 $self->throw_exception("You did not provide any connection_info")
1443 unless defined $info->[0];
1445 my ($old_connect_via, $dbh);
1447 local $DBI::connect_via = 'connect' if $INC{'Apache/DBI.pm'} && $ENV{MOD_PERL};
1449 # this odd anonymous coderef dereference is in fact really
1450 # necessary to avoid the unwanted effect described in perl5
1453 # in addition the coderef itself can't reside inside the try{} block below
1454 # as it somehow triggers a leak under perl -d
1455 my $dbh_error_handler_installer = sub {
1456 weaken (my $weak_self = $_[0]);
1458 # the coderef is blessed so we can distinguish it from externally
1459 # supplied handles (which must be preserved)
1460 $_[1]->{HandleError} = bless sub {
1462 $weak_self->throw_exception("DBI Exception: $_[0]");
1465 # the handler may be invoked by something totally out of
1467 DBIx::Class::Exception->throw("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1469 }, '__DBIC__DBH__ERROR__HANDLER__';
1473 if(ref $info->[0] eq 'CODE') {
1474 $dbh = $info->[0]->();
1478 $dbh = DBI->connect(@$info);
1481 die $DBI::errstr unless $dbh;
1483 die sprintf ("%s fresh DBI handle with a *false* 'Active' attribute. "
1484 . 'This handle is disconnected as far as DBIC is concerned, and we can '
1486 ref $info->[0] eq 'CODE'
1487 ? "Connection coderef $info->[0] returned a"
1488 : 'DBI->connect($schema->storage->connect_info) resulted in a'
1489 ) unless $dbh->FETCH('Active');
1491 # sanity checks unless asked otherwise
1492 unless ($self->unsafe) {
1494 $self->throw_exception(
1495 'Refusing clobbering of {HandleError} installed on externally supplied '
1496 ."DBI handle $dbh. Either remove the handler or use the 'unsafe' attribute."
1497 ) if $dbh->{HandleError} and ref $dbh->{HandleError} ne '__DBIC__DBH__ERROR__HANDLER__';
1499 # Default via _default_dbi_connect_attributes is 1, hence it was an explicit
1500 # request, or an external handle. Complain and set anyway
1501 unless ($dbh->{RaiseError}) {
1502 carp( ref $info->[0] eq 'CODE'
1504 ? "The 'RaiseError' of the externally supplied DBI handle is set to false. "
1505 ."DBIx::Class will toggle it back to true, unless the 'unsafe' connect "
1506 .'attribute has been supplied'
1508 : 'RaiseError => 0 supplied in your connection_info, without an explicit '
1509 .'unsafe => 1. Toggling RaiseError back to true'
1512 $dbh->{RaiseError} = 1;
1515 $dbh_error_handler_installer->($self, $dbh);
1519 $self->throw_exception("DBI Connection failed: $_")
1522 $self->_dbh_autocommit($dbh->{AutoCommit});
1529 # this means we have not yet connected and do not know the AC status
1530 # (e.g. coderef $dbh), need a full-fledged connection check
1531 if (! defined $self->_dbh_autocommit) {
1532 $self->ensure_connected;
1534 # Otherwise simply connect or re-connect on pid changes
1539 $self->next::method(@_);
1542 sub _exec_txn_begin {
1545 # if the user is utilizing txn_do - good for him, otherwise we need to
1546 # ensure that the $dbh is healthy on BEGIN.
1547 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1548 # will be replaced by a failure of begin_work itself (which will be
1549 # then retried on reconnect)
1550 if ($self->{_in_do_block}) {
1551 $self->_dbh->begin_work;
1553 $self->dbh_do(sub { $_[1]->begin_work });
1560 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
1561 $self->throw_exception("Unable to txn_commit() on a disconnected storage")
1564 # esoteric case for folks using external $dbh handles
1565 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1566 carp "Storage transaction_depth 0 does not match "
1567 ."false AutoCommit of $self->{_dbh}, attempting COMMIT anyway";
1568 $self->transaction_depth(1);
1571 $self->next::method(@_);
1573 # if AutoCommit is disabled txn_depth never goes to 0
1574 # as a new txn is started immediately on commit
1575 $self->transaction_depth(1) if (
1576 !$self->transaction_depth
1578 defined $self->_dbh_autocommit
1580 ! $self->_dbh_autocommit
1584 sub _exec_txn_commit {
1585 shift->_dbh->commit;
1591 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
1592 $self->throw_exception("Unable to txn_rollback() on a disconnected storage")
1595 # esoteric case for folks using external $dbh handles
1596 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1597 carp "Storage transaction_depth 0 does not match "
1598 ."false AutoCommit of $self->{_dbh}, attempting ROLLBACK anyway";
1599 $self->transaction_depth(1);
1602 $self->next::method(@_);
1604 # if AutoCommit is disabled txn_depth never goes to 0
1605 # as a new txn is started immediately on commit
1606 $self->transaction_depth(1) if (
1607 !$self->transaction_depth
1609 defined $self->_dbh_autocommit
1611 ! $self->_dbh_autocommit
1615 sub _exec_txn_rollback {
1616 shift->_dbh->rollback;
1619 # generate some identical methods
1620 for my $meth (qw/svp_begin svp_release svp_rollback/) {
1622 *{__PACKAGE__ ."::$meth"} = subname $meth => sub {
1624 $self->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
1625 $self->throw_exception("Unable to $meth() on a disconnected storage")
1627 $self->next::method(@_);
1631 # This used to be the top-half of _execute. It was split out to make it
1632 # easier to override in NoBindVars without duping the rest. It takes up
1633 # all of _execute's args, and emits $sql, @bind.
1634 sub _prep_for_execute {
1635 #my ($self, $op, $ident, $args) = @_;
1636 return shift->_gen_sql_bind(@_)
1640 my ($self, $op, $ident, $args) = @_;
1642 my ($colinfos, $from);
1643 if ( blessed($ident) ) {
1644 $from = $ident->from;
1645 $colinfos = $ident->columns_info;
1649 ($sql, @$bind) = $self->sql_maker->$op( ($from || $ident), @$args );
1651 $bind = $self->_resolve_bindattrs(
1652 $ident, [ @{$args->[2]{bind}||[]}, @$bind ], $colinfos
1656 ! $ENV{DBIC_DT_SEARCH_OK}
1665 $_->[1]->isa('DateTime')
1668 carp_unique 'DateTime objects passed to search() are not supported '
1669 . 'properly (InflateColumn::DateTime formats and settings are not '
1670 . 'respected.) See "Formatting DateTime objects in queries" in '
1671 . 'DBIx::Class::Manual::Cookbook. To disable this warning for good '
1672 . 'set $ENV{DBIC_DT_SEARCH_OK} to true'
1675 return( $sql, $bind );
1678 sub _resolve_bindattrs {
1679 my ($self, $ident, $bind, $colinfos) = @_;
1683 my $resolve_bindinfo = sub {
1684 #my $infohash = shift;
1686 %$colinfos = %{ $self->_resolve_column_info($ident) }
1687 unless keys %$colinfos;
1690 if (my $col = $_[0]->{dbic_colname}) {
1691 $ret = { %{$_[0]} };
1693 $ret->{sqlt_datatype} ||= $colinfos->{$col}{data_type}
1694 if $colinfos->{$col}{data_type};
1696 $ret->{sqlt_size} ||= $colinfos->{$col}{size}
1697 if $colinfos->{$col}{size};
1705 ( ref $_ ne 'ARRAY' or @$_ != 2 ) ? [ {}, $_ ]
1706 : ( ! defined $_->[0] ) ? [ {}, $_->[1] ]
1707 : (ref $_->[0] eq 'HASH') ? [ (exists $_->[0]{dbd_attrs} or $_->[0]{sqlt_datatype})
1709 : $resolve_bindinfo->($_->[0])
1711 : (ref $_->[0] eq 'SCALAR') ? [ { sqlt_datatype => ${$_->[0]} }, $_->[1] ]
1712 : [ $resolve_bindinfo->(
1713 { dbic_colname => $_->[0] }
1718 ! exists $resolved->[0]{dbd_attrs}
1720 ! $resolved->[0]{sqlt_datatype}
1722 length ref $resolved->[1]
1724 ! overload::Method($resolved->[1], '""')
1726 require Data::Dumper;
1727 local $Data::Dumper::Maxdepth = 1;
1728 local $Data::Dumper::Terse = 1;
1729 local $Data::Dumper::Useqq = 1;
1730 local $Data::Dumper::Indent = 0;
1731 local $Data::Dumper::Pad = ' ';
1732 $self->throw_exception(
1733 'You must supply a datatype/bindtype (see DBIx::Class::ResultSet/DBIC BIND VALUES) '
1734 . 'for non-scalar value '. Data::Dumper::Dumper ($resolved->[1])
1743 sub _format_for_trace {
1744 #my ($self, $bind) = @_;
1746 ### Turn @bind from something like this:
1747 ### ( [ "artist", 1 ], [ \%attrs, 3 ] )
1749 ### ( "'1'", "'3'" )
1752 defined( $_ && $_->[1] )
1759 my ( $self, $sql, $bind ) = @_;
1761 $self->debugobj->query_start( $sql, $self->_format_for_trace($bind) )
1766 my ( $self, $sql, $bind ) = @_;
1768 $self->debugobj->query_end( $sql, $self->_format_for_trace($bind) )
1772 sub _dbi_attrs_for_bind {
1773 my ($self, $ident, $bind) = @_;
1777 for (map { $_->[0] } @$bind) {
1779 if (exists $_->{dbd_attrs}) {
1782 elsif($_->{sqlt_datatype}) {
1783 # cache the result in the dbh_details hash, as it can not change unless
1784 # we connect to something else
1785 my $cache = $self->_dbh_details->{_datatype_map_cache} ||= {};
1786 if (not exists $cache->{$_->{sqlt_datatype}}) {
1787 $cache->{$_->{sqlt_datatype}} = $self->bind_attribute_by_data_type($_->{sqlt_datatype}) || undef;
1789 $cache->{$_->{sqlt_datatype}};
1792 undef; # always push something at this position
1801 my ($self, $op, $ident, @args) = @_;
1803 my ($sql, $bind) = $self->_prep_for_execute($op, $ident, \@args);
1805 # not even a PID check - we do not care about the state of the _dbh.
1806 # All we need is to get the appropriate drivers loaded if they aren't
1807 # already so that the assumption in ad7c50fc26e holds
1808 $self->_populate_dbh unless $self->_dbh;
1810 $self->dbh_do( _dbh_execute => # retry over disconnects
1813 $self->_dbi_attrs_for_bind($ident, $bind),
1818 my ($self, $dbh, $sql, $bind, $bind_attrs) = @_;
1820 $self->_query_start( $sql, $bind );
1822 my $sth = $self->_bind_sth_params(
1823 $self->_prepare_sth($dbh, $sql),
1828 # Can this fail without throwing an exception anyways???
1829 my $rv = $sth->execute();
1830 $self->throw_exception(
1831 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1834 $self->_query_end( $sql, $bind );
1836 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1840 my ($self, $dbh, $sql) = @_;
1842 # 3 is the if_active parameter which avoids active sth re-use
1843 my $sth = $self->disable_sth_caching
1844 ? $dbh->prepare($sql)
1845 : $dbh->prepare_cached($sql, {}, 3);
1847 # XXX You would think RaiseError would make this impossible,
1848 # but apparently that's not true :(
1849 $self->throw_exception(
1852 sprintf( "\$dbh->prepare() of '%s' through %s failed *silently* without "
1853 .'an exception and/or setting $dbh->errstr',
1855 ? substr($sql, 0, 20) . '...'
1858 'DBD::' . $dbh->{Driver}{Name},
1865 sub _bind_sth_params {
1866 my ($self, $sth, $bind, $bind_attrs) = @_;
1868 for my $i (0 .. $#$bind) {
1869 if (ref $bind->[$i][1] eq 'SCALAR') { # any scalarrefs are assumed to be bind_inouts
1870 $sth->bind_param_inout(
1871 $i + 1, # bind params counts are 1-based
1873 $bind->[$i][0]{dbd_size} || $self->_max_column_bytesize($bind->[$i][0]), # size
1878 # FIXME SUBOPTIMAL - most likely this is not necessary at all
1879 # confirm with dbi-dev whether explicit stringification is needed
1880 my $v = ( length ref $bind->[$i][1] and overload::Method($bind->[$i][1], '""') )
1895 sub _prefetch_autovalues {
1896 my ($self, $source, $colinfo, $to_insert) = @_;
1899 for my $col (keys %$colinfo) {
1901 $colinfo->{$col}{auto_nextval}
1904 ! exists $to_insert->{$col}
1906 ref $to_insert->{$col} eq 'SCALAR'
1908 (ref $to_insert->{$col} eq 'REF' and ref ${$to_insert->{$col}} eq 'ARRAY')
1911 $values{$col} = $self->_sequence_fetch(
1913 ( $colinfo->{$col}{sequence} ||=
1914 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1924 my ($self, $source, $to_insert) = @_;
1926 my $col_infos = $source->columns_info;
1928 my $prefetched_values = $self->_prefetch_autovalues($source, $col_infos, $to_insert);
1930 # fuse the values, but keep a separate list of prefetched_values so that
1931 # they can be fused once again with the final return
1932 $to_insert = { %$to_insert, %$prefetched_values };
1934 # FIXME - we seem to assume undef values as non-supplied. This is wrong.
1935 # Investigate what does it take to s/defined/exists/
1936 my %pcols = map { $_ => 1 } $source->primary_columns;
1937 my (%retrieve_cols, $autoinc_supplied, $retrieve_autoinc_col);
1938 for my $col ($source->columns) {
1939 if ($col_infos->{$col}{is_auto_increment}) {
1940 $autoinc_supplied ||= 1 if defined $to_insert->{$col};
1941 $retrieve_autoinc_col ||= $col unless $autoinc_supplied;
1944 # nothing to retrieve when explicit values are supplied
1945 next if (defined $to_insert->{$col} and ! (
1946 ref $to_insert->{$col} eq 'SCALAR'
1948 (ref $to_insert->{$col} eq 'REF' and ref ${$to_insert->{$col}} eq 'ARRAY')
1951 # the 'scalar keys' is a trick to preserve the ->columns declaration order
1952 $retrieve_cols{$col} = scalar keys %retrieve_cols if (
1955 $col_infos->{$col}{retrieve_on_insert}
1959 local $self->{_autoinc_supplied_for_op} = $autoinc_supplied;
1960 local $self->{_perform_autoinc_retrieval} = $retrieve_autoinc_col;
1962 my ($sqla_opts, @ir_container);
1963 if (%retrieve_cols and $self->_use_insert_returning) {
1964 $sqla_opts->{returning_container} = \@ir_container
1965 if $self->_use_insert_returning_bound;
1967 $sqla_opts->{returning} = [
1968 sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols
1972 my ($rv, $sth) = $self->_execute('insert', $source, $to_insert, $sqla_opts);
1974 my %returned_cols = %$to_insert;
1975 if (my $retlist = $sqla_opts->{returning}) { # if IR is supported - we will get everything in one set
1976 @ir_container = try {
1977 local $SIG{__WARN__} = sub {};
1978 my @r = $sth->fetchrow_array;
1981 } unless @ir_container;
1983 @returned_cols{@$retlist} = @ir_container if @ir_container;
1986 # pull in PK if needed and then everything else
1987 if (my @missing_pri = grep { $pcols{$_} } keys %retrieve_cols) {
1989 $self->throw_exception( "Missing primary key but Storage doesn't support last_insert_id" )
1990 unless $self->can('last_insert_id');
1992 my @pri_values = $self->last_insert_id($source, @missing_pri);
1994 $self->throw_exception( "Can't get last insert id" )
1995 unless (@pri_values == @missing_pri);
1997 @returned_cols{@missing_pri} = @pri_values;
1998 delete @retrieve_cols{@missing_pri};
2001 # if there is more left to pull
2002 if (%retrieve_cols) {
2003 $self->throw_exception(
2004 'Unable to retrieve additional columns without a Primary Key on ' . $source->source_name
2007 my @left_to_fetch = sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols;
2009 my $cur = DBIx::Class::ResultSet->new($source, {
2010 where => { map { $_ => $returned_cols{$_} } (keys %pcols) },
2011 select => \@left_to_fetch,
2014 @returned_cols{@left_to_fetch} = $cur->next;
2016 $self->throw_exception('Duplicate row returned for PK-search after fresh insert')
2017 if scalar $cur->next;
2021 return { %$prefetched_values, %returned_cols };
2025 my ($self, $source, $cols, $data) = @_;
2027 my @col_range = (0..$#$cols);
2029 # FIXME SUBOPTIMAL - most likely this is not necessary at all
2030 # confirm with dbi-dev whether explicit stringification is needed
2032 # forcibly stringify whatever is stringifiable
2033 # ResultSet::populate() hands us a copy - safe to mangle
2034 for my $r (0 .. $#$data) {
2035 for my $c (0 .. $#{$data->[$r]}) {
2036 $data->[$r][$c] = "$data->[$r][$c]"
2037 if ( length ref $data->[$r][$c] and overload::Method($data->[$r][$c], '""') );
2041 my $colinfos = $source->columns_info($cols);
2043 local $self->{_autoinc_supplied_for_op} =
2044 (first { $_->{is_auto_increment} } values %$colinfos)
2049 # get a slice type index based on first row of data
2050 # a "column" in this context may refer to more than one bind value
2051 # e.g. \[ '?, ?', [...], [...] ]
2053 # construct the value type index - a description of values types for every
2054 # per-column slice of $data:
2056 # nonexistent - nonbind literal
2058 # [] of bindattrs - resolved attribute(s) of bind(s) passed via literal+bind \[] combo
2060 # also construct the column hash to pass to the SQL generator. For plain
2061 # (non literal) values - convert the members of the first row into a
2062 # literal+bind combo, with extra positional info in the bind attr hashref.
2063 # This will allow us to match the order properly, and is so contrived
2064 # because a user-supplied literal/bind (or something else specific to a
2065 # resultsource and/or storage driver) can inject extra binds along the
2066 # way, so one can't rely on "shift positions" ordering at all. Also we
2067 # can't just hand SQLA a set of some known "values" (e.g. hashrefs that
2068 # can be later matched up by address), because we want to supply a real
2069 # value on which perhaps e.g. datatype checks will be performed
2070 my ($proto_data, $value_type_by_col_idx);
2071 for my $i (@col_range) {
2072 my $colname = $cols->[$i];
2073 if (ref $data->[0][$i] eq 'SCALAR') {
2074 # no bind value at all - no type
2076 $proto_data->{$colname} = $data->[0][$i];
2078 elsif (ref $data->[0][$i] eq 'REF' and ref ${$data->[0][$i]} eq 'ARRAY' ) {
2079 # repack, so we don't end up mangling the original \[]
2080 my ($sql, @bind) = @${$data->[0][$i]};
2082 # normalization of user supplied stuff
2083 my $resolved_bind = $self->_resolve_bindattrs(
2084 $source, \@bind, $colinfos,
2087 # store value-less (attrs only) bind info - we will be comparing all
2088 # supplied binds against this for sanity
2089 $value_type_by_col_idx->{$i} = [ map { $_->[0] } @$resolved_bind ];
2091 $proto_data->{$colname} = \[ $sql, map { [
2092 # inject slice order to use for $proto_bind construction
2093 { %{$resolved_bind->[$_][0]}, _bind_data_slice_idx => $i, _literal_bind_subindex => $_+1 }
2095 $resolved_bind->[$_][1]
2100 $value_type_by_col_idx->{$i} = undef;
2102 $proto_data->{$colname} = \[ '?', [
2103 { dbic_colname => $colname, _bind_data_slice_idx => $i }
2110 my ($sql, $proto_bind) = $self->_prep_for_execute (
2116 if (! @$proto_bind and keys %$value_type_by_col_idx) {
2117 # if the bindlist is empty and we had some dynamic binds, this means the
2118 # storage ate them away (e.g. the NoBindVars component) and interpolated
2119 # them directly into the SQL. This obviously can't be good for multi-inserts
2120 $self->throw_exception('Cannot insert_bulk without support for placeholders');
2124 # FIXME - devise a flag "no babysitting" or somesuch to shut this off
2126 # use an error reporting closure for convenience (less to pass)
2127 my $bad_slice_report_cref = sub {
2128 my ($msg, $r_idx, $c_idx) = @_;
2129 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
2133 require Data::Dumper::Concise;
2134 local $Data::Dumper::Maxdepth = 5;
2135 Data::Dumper::Concise::Dumper ({
2136 map { $cols->[$_] =>
2144 for my $col_idx (@col_range) {
2145 my $reference_val = $data->[0][$col_idx];
2147 for my $row_idx (1..$#$data) { # we are comparing against what we got from [0] above, hence start from 1
2148 my $val = $data->[$row_idx][$col_idx];
2150 if (! exists $value_type_by_col_idx->{$col_idx}) { # literal no binds
2151 if (ref $val ne 'SCALAR') {
2152 $bad_slice_report_cref->(
2153 "Incorrect value (expecting SCALAR-ref \\'$$reference_val')",
2158 elsif ($$val ne $$reference_val) {
2159 $bad_slice_report_cref->(
2160 "Inconsistent literal SQL value (expecting \\'$$reference_val')",
2166 elsif (! defined $value_type_by_col_idx->{$col_idx} ) { # regular non-literal value
2167 if (ref $val eq 'SCALAR' or (ref $val eq 'REF' and ref $$val eq 'ARRAY') ) {
2168 $bad_slice_report_cref->("Literal SQL found where a plain bind value is expected", $row_idx, $col_idx);
2171 else { # binds from a \[], compare type and attrs
2172 if (ref $val ne 'REF' or ref $$val ne 'ARRAY') {
2173 $bad_slice_report_cref->(
2174 "Incorrect value (expecting ARRAYREF-ref \\['${$reference_val}->[0]', ... ])",
2179 # start drilling down and bail out early on identical refs
2181 $reference_val != $val
2183 $$reference_val != $$val
2185 if (${$val}->[0] ne ${$reference_val}->[0]) {
2186 $bad_slice_report_cref->(
2187 "Inconsistent literal/bind SQL (expecting \\['${$reference_val}->[0]', ... ])",
2192 # need to check the bind attrs - a bind will happen only once for
2193 # the entire dataset, so any changes further down will be ignored.
2194 elsif (! Data::Compare::Compare(
2195 $value_type_by_col_idx->{$col_idx},
2199 @{$self->_resolve_bindattrs(
2200 $source, [ @{$$val}[1 .. $#$$val] ], $colinfos,
2204 $bad_slice_report_cref->(
2205 'Differing bind attributes on literal/bind values not supported',
2215 # neither _dbh_execute_for_fetch, nor _dbh_execute_inserts_with_no_binds
2216 # are atomic (even if execute_for_fetch is a single call). Thus a safety
2218 my $guard = $self->txn_scope_guard;
2220 $self->_query_start( $sql, @$proto_bind ? [[undef => '__BULK_INSERT__' ]] : () );
2221 my $sth = $self->_prepare_sth($self->_dbh, $sql);
2224 # proto bind contains the information on which pieces of $data to pull
2225 # $cols is passed in only for prettier error-reporting
2226 $self->_dbh_execute_for_fetch( $source, $sth, $proto_bind, $cols, $data );
2229 # bind_param_array doesn't work if there are no binds
2230 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
2234 $self->_query_end( $sql, @$proto_bind ? [[ undef => '__BULK_INSERT__' ]] : () );
2238 return wantarray ? ($rv, $sth, @$proto_bind) : $rv;
2241 # execute_for_fetch is capable of returning data just fine (it means it
2242 # can be used for INSERT...RETURNING and UPDATE...RETURNING. Since this
2243 # is the void-populate fast-path we will just ignore this altogether
2244 # for the time being.
2245 sub _dbh_execute_for_fetch {
2246 my ($self, $source, $sth, $proto_bind, $cols, $data) = @_;
2248 my @idx_range = ( 0 .. $#$proto_bind );
2250 # If we have any bind attributes to take care of, we will bind the
2251 # proto-bind data (which will never be used by execute_for_fetch)
2252 # However since column bindtypes are "sticky", this is sufficient
2253 # to get the DBD to apply the bindtype to all values later on
2255 my $bind_attrs = $self->_dbi_attrs_for_bind($source, $proto_bind);
2257 for my $i (@idx_range) {
2259 $i+1, # DBI bind indexes are 1-based
2260 $proto_bind->[$i][1],
2262 ) if defined $bind_attrs->[$i];
2265 # At this point $data slots named in the _bind_data_slice_idx of
2266 # each piece of $proto_bind are either \[]s or plain values to be
2267 # passed in. Construct the dispensing coderef. *NOTE* the order
2268 # of $data will differ from this of the ?s in the SQL (due to
2269 # alphabetical ordering by colname). We actually do want to
2270 # preserve this behavior so that prepare_cached has a better
2271 # chance of matching on unrelated calls
2273 my $fetch_row_idx = -1; # saner loop this way
2274 my $fetch_tuple = sub {
2275 return undef if ++$fetch_row_idx > $#$data;
2277 return [ map { defined $_->{_literal_bind_subindex}
2278 ? ${ $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]}
2279 ->[ $_->{_literal_bind_subindex} ]
2281 : $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]
2282 } map { $_->[0] } @$proto_bind];
2285 my $tuple_status = [];
2288 $rv = $sth->execute_for_fetch(
2297 # Not all DBDs are create equal. Some throw on error, some return
2298 # an undef $rv, and some set $sth->err - try whatever we can
2299 $err = ($sth->errstr || 'UNKNOWN ERROR ($sth->errstr is unset)') if (
2302 ( !defined $rv or $sth->err )
2305 # Statement must finish even if there was an exception.
2310 $err = shift unless defined $err
2315 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
2317 $self->throw_exception("Unexpected populate error: $err")
2318 if ($i > $#$tuple_status);
2320 require Data::Dumper::Concise;
2321 $self->throw_exception(sprintf "execute_for_fetch() aborted with '%s' at populate slice:\n%s",
2322 ($tuple_status->[$i][1] || $err),
2323 Data::Dumper::Concise::Dumper( { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) } ),
2330 sub _dbh_execute_inserts_with_no_binds {
2331 my ($self, $sth, $count) = @_;
2335 my $dbh = $self->_get_dbh;
2336 local $dbh->{RaiseError} = 1;
2337 local $dbh->{PrintError} = 0;
2339 $sth->execute foreach 1..$count;
2345 # Make sure statement is finished even if there was an exception.
2350 $err = shift unless defined $err;
2353 $self->throw_exception($err) if defined $err;
2359 #my ($self, $source, @args) = @_;
2360 shift->_execute('update', @_);
2365 #my ($self, $source, @args) = @_;
2366 shift->_execute('delete', @_);
2371 $self->_execute($self->_select_args(@_));
2374 sub _select_args_to_query {
2377 $self->throw_exception(
2378 "Unable to generate limited query representation with 'software_limit' enabled"
2379 ) if ($_[3]->{software_limit} and ($_[3]->{offset} or $_[3]->{rows}) );
2381 # my ($op, $ident, $select, $cond, $rs_attrs, $rows, $offset)
2382 # = $self->_select_args($ident, $select, $cond, $attrs);
2383 my ($op, $ident, @args) =
2384 $self->_select_args(@_);
2386 # my ($sql, $prepared_bind) = $self->_gen_sql_bind($op, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
2387 my ($sql, $bind) = $self->_gen_sql_bind($op, $ident, \@args);
2389 # reuse the bind arrayref
2390 unshift @{$bind}, "($sql)";
2395 my ($self, $ident, $select, $where, $orig_attrs) = @_;
2397 # FIXME - that kind of caching would be nice to have
2398 # however currently we *may* pass the same $orig_attrs
2399 # with different ident/select/where
2400 # the whole interface needs to be rethought, since it
2401 # was centered around the flawed SQLA API. We can do
2402 # soooooo much better now. But that is also another
2405 # 'select', @{$orig_attrs->{_sqlmaker_select_args}}
2406 #) if $orig_attrs->{_sqlmaker_select_args};
2408 my $sql_maker = $self->sql_maker;
2409 my $alias2source = $self->_resolve_ident_sources ($ident);
2417 # limit dialects use this stuff
2418 # yes, some CDBICompat crap does not supply an {alias} >.<
2419 ( $orig_attrs->{alias} and $alias2source->{$orig_attrs->{alias}} )
2420 ? ( _rsroot_rsrc => $alias2source->{$orig_attrs->{alias}} )
2425 # Sanity check the attributes (SQLMaker does it too, but
2426 # in case of a software_limit we'll never reach there)
2427 if (defined $attrs->{offset}) {
2428 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
2429 if ( $attrs->{offset} =~ /\D/ or $attrs->{offset} < 0 );
2432 if (defined $attrs->{rows}) {
2433 $self->throw_exception("The rows attribute must be a positive integer if present")
2434 if ( $attrs->{rows} =~ /\D/ or $attrs->{rows} <= 0 );
2436 elsif ($attrs->{offset}) {
2437 # MySQL actually recommends this approach. I cringe.
2438 $attrs->{rows} = $sql_maker->__max_int;
2441 # see if we will need to tear the prefetch apart to satisfy group_by == select
2442 # this is *extremely tricky* to get right, I am still not sure I did
2444 my ($prefetch_needs_subquery, @limit_args);
2446 if ( $attrs->{_grouped_by_distinct} and $attrs->{collapse} ) {
2447 # we already know there is a valid group_by (we made it) and we know it is
2448 # intended to be based *only* on non-multi stuff
2449 # short circuit the group_by parsing below
2450 $prefetch_needs_subquery = 1;
2453 # The rationale is that even if we do *not* have collapse, we still
2454 # need to wrap the core grouped select/group_by in a subquery
2455 # so that databases that care about group_by/select equivalence
2456 # are happy (this includes MySQL in strict_mode)
2457 # If any of the other joined tables are referenced in the group_by
2458 # however - the user is on their own
2459 ( $prefetch_needs_subquery or $attrs->{_related_results_construction} )
2463 @{$attrs->{group_by}}
2465 my $grp_aliases = try { # try{} because $attrs->{from} may be unreadable
2466 $self->_resolve_aliastypes_from_select_args({ from => $attrs->{from}, group_by => $attrs->{group_by} })
2469 # no aliases other than our own in group_by
2470 # if there are - do not allow subquery even if limit is present
2471 $prefetch_needs_subquery = ! scalar grep { $_ ne $attrs->{alias} } keys %{ $grp_aliases->{grouping} || {} };
2473 elsif ( $attrs->{rows} && $attrs->{collapse} ) {
2474 # active collapse with a limit - that one is a no-brainer unless
2475 # overruled by a group_by above
2476 $prefetch_needs_subquery = 1;
2479 if ($prefetch_needs_subquery) {
2480 $attrs = $self->_adjust_select_args_for_complex_prefetch ($attrs);
2482 elsif (! $attrs->{software_limit} ) {
2484 $attrs->{rows} || (),
2485 $attrs->{offset} || (),
2489 # try to simplify the joinmap further (prune unreferenced type-single joins)
2491 ! $prefetch_needs_subquery # already pruned
2495 reftype $attrs->{from} eq 'ARRAY'
2497 @{$attrs->{from}} != 1
2499 ($attrs->{from}, $attrs->{_aliastypes}) = $self->_prune_unused_joins ($attrs);
2503 # This would be the point to deflate anything found in $attrs->{where}
2504 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2505 # expect a result object. And all we have is a resultsource (it is trivial
2506 # to extract deflator coderefs via $alias2source above).
2508 # I don't see a way forward other than changing the way deflators are
2509 # invoked, and that's just bad...
2512 return ( 'select', @{ $orig_attrs->{_sqlmaker_select_args} = [
2513 @{$attrs}{qw(from select where)}, $attrs, @limit_args
2517 # Returns a counting SELECT for a simple count
2518 # query. Abstracted so that a storage could override
2519 # this to { count => 'firstcol' } or whatever makes
2520 # sense as a performance optimization
2522 #my ($self, $source, $rs_attrs) = @_;
2523 return { count => '*' };
2530 =item Arguments: $ident, $select, $condition, $attrs
2534 Handle a SQL select statement.
2540 my ($ident, $select, $condition, $attrs) = @_;
2541 return $self->cursor_class->new($self, \@_, $attrs);
2546 my ($rv, $sth, @bind) = $self->_select(@_);
2547 my @row = $sth->fetchrow_array;
2548 my @nextrow = $sth->fetchrow_array if @row;
2549 if(@row && @nextrow) {
2550 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2552 # Need to call finish() to work round broken DBDs
2557 =head2 sql_limit_dialect
2559 This is an accessor for the default SQL limit dialect used by a particular
2560 storage driver. Can be overridden by supplying an explicit L</limit_dialect>
2561 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2562 see L<DBIx::Class::SQLMaker::LimitDialects>.
2566 sub _dbh_columns_info_for {
2567 my ($self, $dbh, $table) = @_;
2569 if ($dbh->can('column_info')) {
2573 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2574 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2576 while ( my $info = $sth->fetchrow_hashref() ){
2578 $column_info{data_type} = $info->{TYPE_NAME};
2579 $column_info{size} = $info->{COLUMN_SIZE};
2580 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2581 $column_info{default_value} = $info->{COLUMN_DEF};
2582 my $col_name = $info->{COLUMN_NAME};
2583 $col_name =~ s/^\"(.*)\"$/$1/;
2585 $result{$col_name} = \%column_info;
2590 return \%result if !$caught && scalar keys %result;
2594 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2596 my @columns = @{$sth->{NAME_lc}};
2597 for my $i ( 0 .. $#columns ){
2599 $column_info{data_type} = $sth->{TYPE}->[$i];
2600 $column_info{size} = $sth->{PRECISION}->[$i];
2601 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2603 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2604 $column_info{data_type} = $1;
2605 $column_info{size} = $2;
2608 $result{$columns[$i]} = \%column_info;
2612 foreach my $col (keys %result) {
2613 my $colinfo = $result{$col};
2614 my $type_num = $colinfo->{data_type};
2616 if(defined $type_num && $dbh->can('type_info')) {
2617 my $type_info = $dbh->type_info($type_num);
2618 $type_name = $type_info->{TYPE_NAME} if $type_info;
2619 $colinfo->{data_type} = $type_name if $type_name;
2626 sub columns_info_for {
2627 my ($self, $table) = @_;
2628 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2631 =head2 last_insert_id
2633 Return the row id of the last insert.
2637 sub _dbh_last_insert_id {
2638 my ($self, $dbh, $source, $col) = @_;
2640 my $id = try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2642 return $id if defined $id;
2644 my $class = ref $self;
2645 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2648 sub last_insert_id {
2650 $self->_dbh_last_insert_id ($self->_dbh, @_);
2653 =head2 _native_data_type
2657 =item Arguments: $type_name
2661 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2662 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2663 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2665 The default implementation returns C<undef>, implement in your Storage driver if
2666 you need this functionality.
2668 Should map types from other databases to the native RDBMS type, for example
2669 C<VARCHAR2> to C<VARCHAR>.
2671 Types with modifiers should map to the underlying data type. For example,
2672 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2674 Composite types should map to the container type, for example
2675 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2679 sub _native_data_type {
2680 #my ($self, $data_type) = @_;
2684 # Check if placeholders are supported at all
2685 sub _determine_supports_placeholders {
2687 my $dbh = $self->_get_dbh;
2689 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2690 # but it is inaccurate more often than not
2692 local $dbh->{PrintError} = 0;
2693 local $dbh->{RaiseError} = 1;
2694 $dbh->do('select ?', {}, 1);
2702 # Check if placeholders bound to non-string types throw exceptions
2704 sub _determine_supports_typeless_placeholders {
2706 my $dbh = $self->_get_dbh;
2709 local $dbh->{PrintError} = 0;
2710 local $dbh->{RaiseError} = 1;
2711 # this specifically tests a bind that is NOT a string
2712 $dbh->do('select 1 where 1 = ?', {}, 1);
2722 Returns the database driver name.
2727 shift->_get_dbh->{Driver}->{Name};
2730 =head2 bind_attribute_by_data_type
2732 Given a datatype from column info, returns a database specific bind
2733 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2734 let the database planner just handle it.
2736 This method is always called after the driver has been determined and a DBI
2737 connection has been established. Therefore you can refer to C<DBI::$constant>
2738 and/or C<DBD::$driver::$constant> directly, without worrying about loading
2739 the correct modules.
2743 sub bind_attribute_by_data_type {
2747 =head2 is_datatype_numeric
2749 Given a datatype from column_info, returns a boolean value indicating if
2750 the current RDBMS considers it a numeric value. This controls how
2751 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2752 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2753 be performed instead of the usual C<eq>.
2757 sub is_datatype_numeric {
2758 #my ($self, $dt) = @_;
2760 return 0 unless $_[1];
2763 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2768 =head2 create_ddl_dir
2772 =item Arguments: $schema, \@databases, $version, $directory, $preversion, \%sqlt_args
2776 Creates a SQL file based on the Schema, for each of the specified
2777 database engines in C<\@databases> in the given directory.
2778 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2780 Given a previous version number, this will also create a file containing
2781 the ALTER TABLE statements to transform the previous schema into the
2782 current one. Note that these statements may contain C<DROP TABLE> or
2783 C<DROP COLUMN> statements that can potentially destroy data.
2785 The file names are created using the C<ddl_filename> method below, please
2786 override this method in your schema if you would like a different file
2787 name format. For the ALTER file, the same format is used, replacing
2788 $version in the name with "$preversion-$version".
2790 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2791 The most common value for this would be C<< { add_drop_table => 1 } >>
2792 to have the SQL produced include a C<DROP TABLE> statement for each table
2793 created. For quoting purposes supply C<quote_identifiers>.
2795 If no arguments are passed, then the following default values are assumed:
2799 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2801 =item version - $schema->schema_version
2803 =item directory - './'
2805 =item preversion - <none>
2809 By default, C<\%sqlt_args> will have
2811 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2813 merged with the hash passed in. To disable any of those features, pass in a
2814 hashref like the following
2816 { ignore_constraint_names => 0, # ... other options }
2819 WARNING: You are strongly advised to check all SQL files created, before applying
2824 sub create_ddl_dir {
2825 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2828 carp "No directory given, using ./\n";
2833 (require File::Path and File::Path::mkpath (["$dir"])) # mkpath does not like objects (i.e. Path::Class::Dir)
2835 $self->throw_exception(
2836 "Failed to create '$dir': " . ($! || $@ || 'error unknown')
2840 $self->throw_exception ("Directory '$dir' does not exist\n") unless(-d $dir);
2842 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2843 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2845 my $schema_version = $schema->schema_version || '1.x';
2846 $version ||= $schema_version;
2849 add_drop_table => 1,
2850 ignore_constraint_names => 1,
2851 ignore_index_names => 1,
2855 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy')) {
2856 $self->throw_exception("Can't create a ddl file without " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2859 my $sqlt = SQL::Translator->new( $sqltargs );
2861 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2862 my $sqlt_schema = $sqlt->translate({ data => $schema })
2863 or $self->throw_exception ($sqlt->error);
2865 foreach my $db (@$databases) {
2867 $sqlt->{schema} = $sqlt_schema;
2868 $sqlt->producer($db);
2871 my $filename = $schema->ddl_filename($db, $version, $dir);
2872 if (-e $filename && ($version eq $schema_version )) {
2873 # if we are dumping the current version, overwrite the DDL
2874 carp "Overwriting existing DDL file - $filename";
2878 my $output = $sqlt->translate;
2880 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2883 if(!open($file, ">$filename")) {
2884 $self->throw_exception("Can't open $filename for writing ($!)");
2887 print $file $output;
2890 next unless ($preversion);
2892 require SQL::Translator::Diff;
2894 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2895 if(!-e $prefilename) {
2896 carp("No previous schema file found ($prefilename)");
2900 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2902 carp("Overwriting existing diff file - $difffile");
2908 my $t = SQL::Translator->new($sqltargs);
2913 or $self->throw_exception ($t->error);
2915 my $out = $t->translate( $prefilename )
2916 or $self->throw_exception ($t->error);
2918 $source_schema = $t->schema;
2920 $source_schema->name( $prefilename )
2921 unless ( $source_schema->name );
2924 # The "new" style of producers have sane normalization and can support
2925 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2926 # And we have to diff parsed SQL against parsed SQL.
2927 my $dest_schema = $sqlt_schema;
2929 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2930 my $t = SQL::Translator->new($sqltargs);
2935 or $self->throw_exception ($t->error);
2937 my $out = $t->translate( $filename )
2938 or $self->throw_exception ($t->error);
2940 $dest_schema = $t->schema;
2942 $dest_schema->name( $filename )
2943 unless $dest_schema->name;
2946 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2950 if(!open $file, ">$difffile") {
2951 $self->throw_exception("Can't write to $difffile ($!)");
2959 =head2 deployment_statements
2963 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2967 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2969 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2970 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2972 C<$directory> is used to return statements from files in a previously created
2973 L</create_ddl_dir> directory and is optional. The filenames are constructed
2974 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2976 If no C<$directory> is specified then the statements are constructed on the
2977 fly using L<SQL::Translator> and C<$version> is ignored.
2979 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2983 sub deployment_statements {
2984 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2985 $type ||= $self->sqlt_type;
2986 $version ||= $schema->schema_version || '1.x';
2988 my $filename = $schema->ddl_filename($type, $version, $dir);
2991 # FIXME replace this block when a proper sane sql parser is available
2993 open($file, "<$filename")
2994 or $self->throw_exception("Can't open $filename ($!)");
2997 return join('', @rows);
3000 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy') ) {
3001 $self->throw_exception("Can't deploy without a ddl_dir or " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
3004 # sources needs to be a parser arg, but for simplicity allow at top level
3006 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
3007 if exists $sqltargs->{sources};
3009 my $tr = SQL::Translator->new(
3010 producer => "SQL::Translator::Producer::${type}",
3012 parser => 'SQL::Translator::Parser::DBIx::Class',
3016 return preserve_context {
3019 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
3020 unless defined $_[0];
3024 # FIXME deploy() currently does not accurately report sql errors
3025 # Will always return true while errors are warned
3027 my ($self, $schema, $type, $sqltargs, $dir) = @_;
3031 return if($line =~ /^--/);
3032 # next if($line =~ /^DROP/m);
3033 return if($line =~ /^BEGIN TRANSACTION/m);
3034 return if($line =~ /^COMMIT/m);
3035 return if $line =~ /^\s+$/; # skip whitespace only
3036 $self->_query_start($line);
3038 # do a dbh_do cycle here, as we need some error checking in
3039 # place (even though we will ignore errors)
3040 $self->dbh_do (sub { $_[1]->do($line) });
3042 carp qq{$_ (running "${line}")};
3044 $self->_query_end($line);
3046 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
3047 if (@statements > 1) {
3048 foreach my $statement (@statements) {
3049 $deploy->( $statement );
3052 elsif (@statements == 1) {
3053 # split on single line comments and end of statements
3054 foreach my $line ( split(/\s*--.*\n|;\n/, $statements[0])) {
3060 =head2 datetime_parser
3062 Returns the datetime parser class
3066 sub datetime_parser {
3068 return $self->{datetime_parser} ||= do {
3069 $self->build_datetime_parser(@_);
3073 =head2 datetime_parser_type
3075 Defines the datetime parser class - currently defaults to L<DateTime::Format::MySQL>
3077 =head2 build_datetime_parser
3079 See L</datetime_parser>
3083 sub build_datetime_parser {
3085 my $type = $self->datetime_parser_type(@_);
3090 =head2 is_replicating
3092 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
3093 replicate from a master database. Default is undef, which is the result
3094 returned by databases that don't support replication.
3098 sub is_replicating {
3103 =head2 lag_behind_master
3105 Returns a number that represents a certain amount of lag behind a master db
3106 when a given storage is replicating. The number is database dependent, but
3107 starts at zero and increases with the amount of lag. Default in undef
3111 sub lag_behind_master {
3115 =head2 relname_to_table_alias
3119 =item Arguments: $relname, $join_count
3121 =item Return Value: $alias
3125 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
3128 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
3129 way these aliases are named.
3131 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
3132 otherwise C<"$relname">.
3136 sub relname_to_table_alias {
3137 my ($self, $relname, $join_count) = @_;
3139 my $alias = ($join_count && $join_count > 1 ?
3140 join('_', $relname, $join_count) : $relname);
3145 # The size in bytes to use for DBI's ->bind_param_inout, this is the generic
3146 # version and it may be necessary to amend or override it for a specific storage
3147 # if such binds are necessary.
3148 sub _max_column_bytesize {
3149 my ($self, $attr) = @_;
3153 if ($attr->{sqlt_datatype}) {
3154 my $data_type = lc($attr->{sqlt_datatype});
3156 if ($attr->{sqlt_size}) {
3158 # String/sized-binary types
3159 if ($data_type =~ /^(?:
3160 l? (?:var)? char(?:acter)? (?:\s*varying)?
3162 (?:var)? binary (?:\s*varying)?
3167 $max_size = $attr->{sqlt_size};
3169 # Other charset/unicode types, assume scale of 4
3170 elsif ($data_type =~ /^(?:
3171 national \s* character (?:\s*varying)?
3180 $max_size = $attr->{sqlt_size} * 4;
3184 if (!$max_size and !$self->_is_lob_type($data_type)) {
3185 $max_size = 100 # for all other (numeric?) datatypes
3189 $max_size || $self->_dbic_connect_attributes->{LongReadLen} || $self->_get_dbh->{LongReadLen} || 8000;
3192 # Determine if a data_type is some type of BLOB
3194 my ($self, $data_type) = @_;
3195 $data_type && ($data_type =~ /lob|bfile|text|image|bytea|memo/i
3196 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary
3197 |varchar|character\s*varying|nvarchar
3198 |national\s*character\s*varying))?\z/xi);
3201 sub _is_binary_lob_type {
3202 my ($self, $data_type) = @_;
3203 $data_type && ($data_type =~ /blob|bfile|image|bytea/i
3204 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary))?\z/xi);
3207 sub _is_text_lob_type {
3208 my ($self, $data_type) = @_;
3209 $data_type && ($data_type =~ /^(?:clob|memo)\z/i
3210 || $data_type =~ /^long(?:\s+(?:varchar|character\s*varying|nvarchar
3211 |national\s*character\s*varying))\z/xi);
3214 # Determine if a data_type is some type of a binary type
3215 sub _is_binary_type {
3216 my ($self, $data_type) = @_;
3217 $data_type && ($self->_is_binary_lob_type($data_type)
3218 || $data_type =~ /(?:var)?(?:binary|bit|graphic)(?:\s*varying)?/i);
3225 =head2 DBIx::Class and AutoCommit
3227 DBIx::Class can do some wonderful magic with handling exceptions,
3228 disconnections, and transactions when you use C<< AutoCommit => 1 >>
3229 (the default) combined with L<txn_do|DBIx::Class::Storage/txn_do> for
3230 transaction support.
3232 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
3233 in an assumed transaction between commits, and you're telling us you'd
3234 like to manage that manually. A lot of the magic protections offered by
3235 this module will go away. We can't protect you from exceptions due to database
3236 disconnects because we don't know anything about how to restart your
3237 transactions. You're on your own for handling all sorts of exceptional
3238 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
3242 =head1 AUTHOR AND CONTRIBUTORS
3244 See L<AUTHOR|DBIx::Class/AUTHOR> and L<CONTRIBUTORS|DBIx::Class/CONTRIBUTORS> in DBIx::Class
3248 You may distribute this code under the same terms as Perl itself.