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 Context::Preserve 'preserve_context';
15 use SQL::Abstract qw(is_plain_value is_literal_value);
16 use DBIx::Class::_Util qw(
17 quote_sub perlstring serialize
19 detected_reinvoked_destructor scope_guard
23 # default cursor class, overridable in connect_info attributes
24 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
26 __PACKAGE__->mk_group_accessors('inherited' => qw/
27 sql_limit_dialect sql_quote_char sql_name_sep
30 __PACKAGE__->mk_group_accessors('component_class' => qw/sql_maker_class datetime_parser_type/);
32 __PACKAGE__->sql_maker_class('DBIx::Class::SQLMaker');
33 __PACKAGE__->datetime_parser_type('DateTime::Format::MySQL'); # historic default
35 __PACKAGE__->sql_name_sep('.');
37 __PACKAGE__->mk_group_accessors('simple' => qw/
38 _connect_info _dbic_connect_attributes _driver_determined
39 _dbh _dbh_details _conn_pid _sql_maker _sql_maker_opts _dbh_autocommit
40 _perform_autoinc_retrieval _autoinc_supplied_for_op
43 # the values for these accessors are picked out (and deleted) from
44 # the attribute hashref passed to connect_info
45 my @storage_options = qw/
46 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
47 disable_sth_caching unsafe auto_savepoint
49 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
52 # capability definitions, using a 2-tiered accessor system
55 # A driver/user may define _use_X, which blindly without any checks says:
56 # "(do not) use this capability", (use_dbms_capability is an "inherited"
59 # If _use_X is undef, _supports_X is then queried. This is a "simple" style
60 # accessor, which in turn calls _determine_supports_X, and stores the return
61 # in a special slot on the storage object, which is wiped every time a $dbh
62 # reconnection takes place (it is not guaranteed that upon reconnection we
63 # will get the same rdbms version). _determine_supports_X does not need to
64 # exist on a driver, as we ->can for it before calling.
66 my @capabilities = (qw/
68 insert_returning_bound
77 __PACKAGE__->mk_group_accessors( dbms_capability => map { "_supports_$_" } @capabilities );
78 __PACKAGE__->mk_group_accessors( use_dbms_capability => map { "_use_$_" } (@capabilities ) );
80 # on by default, not strictly a capability (pending rewrite)
81 __PACKAGE__->_use_join_optimizer (1);
82 sub _determine_supports_join_optimizer { 1 };
84 # Each of these methods need _determine_driver called before itself
85 # in order to function reliably. We also need to separate accessors
86 # from plain old method calls, since an accessor called as a setter
87 # does *not* need the driver determination loop fired (and in fact
88 # can produce hard to find bugs, like e.g. losing on_connect_*
89 # semantics on fresh connections)
91 # The construct below is simply a parameterized around()
92 my $storage_accessor_idx = { map { $_ => 1 } qw(
99 for my $meth (keys %$storage_accessor_idx, qw(
100 deployment_statements
102 build_datetime_parser
114 with_deferred_fk_checks
116 get_use_dbms_capability
123 my $orig = __PACKAGE__->can ($meth)
124 or die "$meth is not a ::Storage::DBI method!";
126 my $possibly_a_setter = $storage_accessor_idx->{$meth} ? 1 : 0;
129 __PACKAGE__ ."::$meth", sprintf( <<'EOC', $possibly_a_setter, perlstring $meth ), { '$orig' => \$orig };
132 # if this is an actual *setter* - just set it, no need to connect
133 # and determine the driver
136 # only fire when invoked on an instance, a valid class-based invocation
137 # would e.g. be setting a default for an inherited accessor
140 ! $_[0]->{_driver_determined}
142 ! $_[0]->{_in_determine_driver}
144 # Only try to determine stuff if we have *something* that either is or can
145 # provide a DSN. Allows for bare $schema's generated with a plain ->connect()
146 # to still be marginally useful
147 $_[0]->_dbi_connect_info->[0]
149 $_[0]->_determine_driver;
151 # work around http://rt.perl.org/rt3/Public/Bug/Display.html?id=35878
152 goto $_[0]->can(%2$s) unless DBIx::Class::_ENV_::BROKEN_GOTO;
154 my $cref = $_[0]->can(%2$s);
164 DBIx::Class::Storage::DBI - DBI storage handler
168 my $schema = MySchema->connect('dbi:SQLite:my.db');
170 $schema->storage->debug(1);
172 my @stuff = $schema->storage->dbh_do(
174 my ($storage, $dbh, @args) = @_;
175 $dbh->do("DROP TABLE authors");
180 $schema->resultset('Book')->search({
181 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
186 This class represents the connection to an RDBMS via L<DBI>. See
187 L<DBIx::Class::Storage> for general information. This pod only
188 documents DBI-specific methods and behaviors.
195 my $new = shift->next::method(@_);
197 $new->_sql_maker_opts({});
198 $new->_dbh_details({});
199 $new->{_in_do_block} = 0;
201 # read below to see what this does
202 $new->_arm_global_destructor;
207 # This is hack to work around perl shooting stuff in random
208 # order on exit(). If we do not walk the remaining storage
209 # objects in an END block, there is a *small but real* chance
210 # of a fork()ed child to kill the parent's shared DBI handle,
211 # *before perl reaches the DESTROY in this package*
212 # Yes, it is ugly and effective.
213 # Additionally this registry is used by the CLONE method to
214 # make sure no handles are shared between threads
216 my %seek_and_destroy;
218 sub _arm_global_destructor {
220 # quick "garbage collection" pass - prevents the registry
221 # from slowly growing with a bunch of undef-valued keys
222 defined $seek_and_destroy{$_} or delete $seek_and_destroy{$_}
223 for keys %seek_and_destroy;
226 $seek_and_destroy{ refaddr($_[0]) } = $_[0]
229 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
230 # collected before leaving this scope. Depending on the code above, this
231 # may very well be just a preventive measure guarding future modifications
238 ! DBIx::Class::_ENV_::BROKEN_FORK
240 my @instances = grep { defined $_ } values %seek_and_destroy
242 local $?; # just in case the DBI destructor changes it somehow
244 # disarm the handle if not native to this process (see comment on top)
245 $_->_verify_pid for @instances;
248 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
249 # collected before leaving this scope. Depending on the code above, this
250 # may very well be just a preventive measure guarding future modifications
255 # As per DBI's recommendation, DBIC disconnects all handles as
256 # soon as possible (DBIC will reconnect only on demand from within
258 my @instances = grep { defined $_ } values %seek_and_destroy;
259 %seek_and_destroy = ();
265 # properly renumber existing refs
266 $_->_arm_global_destructor
269 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
270 # collected before leaving this scope. Depending on the code above, this
271 # may very well be just a preventive measure guarding future modifications
277 return if &detected_reinvoked_destructor;
279 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
281 # some databases spew warnings on implicit disconnect
282 return unless defined $_[0]->_dbh;
284 local $SIG{__WARN__} = sub {};
286 # not calling ->disconnect here - we are being destroyed - nothing to reset
288 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
289 # collected before leaving this scope. Depending on the code above, this
290 # may very well be just a preventive measure guarding future modifications
294 # handle pid changes correctly - do not destroy parent's connection
297 my $pid = $_[0]->_conn_pid;
299 if( defined $pid and $pid != $$ and my $dbh = $_[0]->_dbh ) {
300 $dbh->{InactiveDestroy} = 1;
305 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
306 # collected before leaving this scope. Depending on the code above, this
307 # may very well be just a preventive measure guarding future modifications
313 This method is normally called by L<DBIx::Class::Schema/connection>, which
314 encapsulates its argument list in an arrayref before passing them here.
316 The argument list may contain:
322 The same 4-element argument set one would normally pass to
323 L<DBI/connect>, optionally followed by
324 L<extra attributes|/DBIx::Class specific connection attributes>
325 recognized by DBIx::Class:
327 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
331 A single code reference which returns a connected
332 L<DBI database handle|DBI/connect> optionally followed by
333 L<extra attributes|/DBIx::Class specific connection attributes> recognized
336 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
340 A single hashref with all the attributes and the dsn/user/password
343 $connect_info_args = [{
351 $connect_info_args = [{
352 dbh_maker => sub { DBI->connect (...) },
357 This is particularly useful for L<Catalyst> based applications, allowing the
358 following config (L<Config::General> style):
363 dsn dbi:mysql:database=test
370 The C<dsn>/C<user>/C<password> combination can be substituted by the
371 C<dbh_maker> key whose value is a coderef that returns a connected
372 L<DBI database handle|DBI/connect>
376 Please note that the L<DBI> docs recommend that you always explicitly
377 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
378 recommends that it be set to I<1>, and that you perform transactions
379 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
380 to I<1> if you do not do explicitly set it to zero. This is the default
381 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
383 =head3 DBIx::Class specific connection attributes
385 In addition to the standard L<DBI|DBI/ATTRIBUTES COMMON TO ALL HANDLES>
386 L<connection|DBI/Database Handle Attributes> attributes, DBIx::Class recognizes
387 the following connection options. These options can be mixed in with your other
388 L<DBI> connection attributes, or placed in a separate hashref
389 (C<\%extra_attributes>) as shown above.
391 Every time C<connect_info> is invoked, any previous settings for
392 these options will be cleared before setting the new ones, regardless of
393 whether any options are specified in the new C<connect_info>.
400 Specifies things to do immediately after connecting or re-connecting to
401 the database. Its value may contain:
407 This contains one SQL statement to execute.
409 =item an array reference
411 This contains SQL statements to execute in order. Each element contains
412 a string or a code reference that returns a string.
414 =item a code reference
416 This contains some code to execute. Unlike code references within an
417 array reference, its return value is ignored.
421 =item on_disconnect_do
423 Takes arguments in the same form as L</on_connect_do> and executes them
424 immediately before disconnecting from the database.
426 Note, this only runs if you explicitly call L</disconnect> on the
429 =item on_connect_call
431 A more generalized form of L</on_connect_do> that calls the specified
432 C<connect_call_METHOD> methods in your storage driver.
434 on_connect_do => 'select 1'
438 on_connect_call => [ [ do_sql => 'select 1' ] ]
440 Its values may contain:
446 Will call the C<connect_call_METHOD> method.
448 =item a code reference
450 Will execute C<< $code->($storage) >>
452 =item an array reference
454 Each value can be a method name or code reference.
456 =item an array of arrays
458 For each array, the first item is taken to be the C<connect_call_> method name
459 or code reference, and the rest are parameters to it.
463 Some predefined storage methods you may use:
469 Executes a SQL string or a code reference that returns a SQL string. This is
470 what L</on_connect_do> and L</on_disconnect_do> use.
478 Will execute the scalar as SQL.
482 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
483 attributes hashref and bind values.
485 =item a code reference
487 Will execute C<< $code->($storage) >> and execute the return array refs as
494 Execute any statements necessary to initialize the database session to return
495 and accept datetime/timestamp values used with
496 L<DBIx::Class::InflateColumn::DateTime>.
498 Only necessary for some databases, see your specific storage driver for
499 implementation details.
503 =item on_disconnect_call
505 Takes arguments in the same form as L</on_connect_call> and executes them
506 immediately before disconnecting from the database.
508 Calls the C<disconnect_call_METHOD> methods as opposed to the
509 C<connect_call_METHOD> methods called by L</on_connect_call>.
511 Note, this only runs if you explicitly call L</disconnect> on the
514 =item disable_sth_caching
516 If set to a true value, this option will disable the caching of
517 statement handles via L<DBI/prepare_cached>.
521 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
522 default L</sql_limit_dialect> setting of the storage (if any). For a list
523 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
527 When true automatically sets L</quote_char> and L</name_sep> to the characters
528 appropriate for your particular RDBMS. This option is preferred over specifying
529 L</quote_char> directly.
533 Specifies what characters to use to quote table and column names.
535 C<quote_char> expects either a single character, in which case is it
536 is placed on either side of the table/column name, or an arrayref of length
537 2 in which case the table/column name is placed between the elements.
539 For example under MySQL you should use C<< quote_char => '`' >>, and for
540 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
544 This parameter is only useful in conjunction with C<quote_char>, and is used to
545 specify the character that separates elements (schemas, tables, columns) from
546 each other. If unspecified it defaults to the most commonly used C<.>.
550 This Storage driver normally installs its own C<HandleError>, sets
551 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
552 all database handles, including those supplied by a coderef. It does this
553 so that it can have consistent and useful error behavior.
555 If you set this option to a true value, Storage will not do its usual
556 modifications to the database handle's attributes, and instead relies on
557 the settings in your connect_info DBI options (or the values you set in
558 your connection coderef, in the case that you are connecting via coderef).
560 Note that your custom settings can cause Storage to malfunction,
561 especially if you set a C<HandleError> handler that suppresses exceptions
562 and/or disable C<RaiseError>.
566 If this option is true, L<DBIx::Class> will use savepoints when nesting
567 transactions, making it possible to recover from failure in the inner
568 transaction without having to abort all outer transactions.
572 Use this argument to supply a cursor class other than the default
573 L<DBIx::Class::Storage::DBI::Cursor>.
577 Some real-life examples of arguments to L</connect_info> and
578 L<DBIx::Class::Schema/connect>
580 # Simple SQLite connection
581 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
584 ->connect_info([ sub { DBI->connect(...) } ]);
586 # Connect via subref in hashref
588 dbh_maker => sub { DBI->connect(...) },
589 on_connect_do => 'alter session ...',
592 # A bit more complicated
599 { quote_char => q{"} },
603 # Equivalent to the previous example
609 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
613 # Same, but with hashref as argument
614 # See parse_connect_info for explanation
617 dsn => 'dbi:Pg:dbname=foo',
619 password => 'my_pg_password',
626 # Subref + DBIx::Class-specific connection options
629 sub { DBI->connect(...) },
633 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
634 disable_sth_caching => 1,
644 my ($self, $info) = @_;
646 return $self->_connect_info if !$info;
648 $self->_connect_info($info); # copy for _connect_info
650 $info = $self->_normalize_connect_info($info)
651 if ref $info eq 'ARRAY';
654 %{ $self->_default_dbi_connect_attributes || {} },
655 %{ $info->{attributes} || {} },
658 my @args = @{ $info->{arguments} };
660 if (keys %attrs and ref $args[0] ne 'CODE') {
662 'You provided explicit AutoCommit => 0 in your connection_info. '
663 . 'This is almost universally a bad idea (see the footnotes of '
664 . 'DBIx::Class::Storage::DBI for more info). If you still want to '
665 . 'do this you can set $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK} to disable '
667 ) if ! $attrs{AutoCommit} and ! $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK};
669 push @args, \%attrs if keys %attrs;
672 # this is the authoritative "always an arrayref" thing fed to DBI->connect
673 # OR a single-element coderef-based $dbh factory
674 $self->_dbi_connect_info(\@args);
676 # extract the individual storage options
677 for my $storage_opt (keys %{ $info->{storage_options} }) {
678 my $value = $info->{storage_options}{$storage_opt};
680 $self->$storage_opt($value);
683 # Extract the individual sqlmaker options
685 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
686 # the new set of options
687 $self->_sql_maker(undef);
688 $self->_sql_maker_opts({});
690 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
691 my $value = $info->{sql_maker_options}{$sql_maker_opt};
693 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
697 # save attributes in a separate accessor so they are always
698 # introspectable, even in case of a CODE $dbhmaker
699 $self->_dbic_connect_attributes (\%attrs);
701 return $self->_connect_info;
704 sub _dbi_connect_info {
707 return $self->{_dbi_connect_info} = $_[0]
710 my $conninfo = $self->{_dbi_connect_info} || [];
712 # last ditch effort to grab a DSN
713 if ( ! defined $conninfo->[0] and $ENV{DBI_DSN} ) {
714 my @new_conninfo = @$conninfo;
715 $new_conninfo[0] = $ENV{DBI_DSN};
716 $conninfo = \@new_conninfo;
723 sub _normalize_connect_info {
724 my ($self, $info_arg) = @_;
727 my @args = @$info_arg; # take a shallow copy for further mutilation
729 # combine/pre-parse arguments depending on invocation style
732 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
733 %attrs = %{ $args[1] || {} };
736 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
737 %attrs = %{$args[0]};
739 if (my $code = delete $attrs{dbh_maker}) {
742 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
745 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
746 . "to the result of 'dbh_maker'",
748 join (', ', map { "'$_'" } (@ignored) ),
753 @args = delete @attrs{qw/dsn user password/};
756 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
758 % { $args[3] || {} },
759 % { $args[4] || {} },
761 @args = @args[0,1,2];
764 $info{arguments} = \@args;
766 my @storage_opts = grep exists $attrs{$_},
767 @storage_options, 'cursor_class';
769 @{ $info{storage_options} }{@storage_opts} =
770 delete @attrs{@storage_opts} if @storage_opts;
772 my @sql_maker_opts = grep exists $attrs{$_},
773 qw/limit_dialect quote_char name_sep quote_names/;
775 @{ $info{sql_maker_options} }{@sql_maker_opts} =
776 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
778 $info{attributes} = \%attrs if %attrs;
783 sub _default_dbi_connect_attributes () {
788 ShowErrorStatement => 1,
794 This method is deprecated in favour of setting via L</connect_info>.
798 =head2 on_disconnect_do
800 This method is deprecated in favour of setting via L</connect_info>.
804 sub _parse_connect_do {
805 my ($self, $type) = @_;
807 my $val = $self->$type;
808 return () if not defined $val;
813 push @res, [ 'do_sql', $val ];
814 } elsif (ref($val) eq 'CODE') {
816 } elsif (ref($val) eq 'ARRAY') {
817 push @res, map { [ 'do_sql', $_ ] } @$val;
819 $self->throw_exception("Invalid type for $type: ".ref($val));
827 Arguments: ($subref | $method_name), @extra_coderef_args?
829 Execute the given $subref or $method_name using the new exception-based
830 connection management.
832 The first two arguments will be the storage object that C<dbh_do> was called
833 on and a database handle to use. Any additional arguments will be passed
834 verbatim to the called subref as arguments 2 and onwards.
836 Using this (instead of $self->_dbh or $self->dbh) ensures correct
837 exception handling and reconnection (or failover in future subclasses).
839 Your subref should have no side-effects outside of the database, as
840 there is the potential for your subref to be partially double-executed
841 if the database connection was stale/dysfunctional.
845 my @stuff = $schema->storage->dbh_do(
847 my ($storage, $dbh, @cols) = @_;
848 my $cols = join(q{, }, @cols);
849 $dbh->selectrow_array("SELECT $cols FROM foo");
858 my $run_target = shift; # either a coderef or a method name
860 # short circuit when we know there is no need for a runner
862 # FIXME - assumption may be wrong
863 # the rationale for the txn_depth check is that if this block is a part
864 # of a larger transaction, everything up to that point is screwed anyway
865 return $self->$run_target($self->_get_dbh, @_)
866 if $self->{_in_do_block} or $self->transaction_depth;
868 # take a ref instead of a copy, to preserve @_ aliasing
869 # semantics within the coderef, but only if needed
870 # (pseudoforking doesn't like this trick much)
871 my $args = @_ ? \@_ : [];
873 DBIx::Class::Storage::BlockRunner->new(
876 retry_handler => sub {
877 $_[0]->failed_attempt_count == 1
879 ! $_[0]->storage->connected
882 $self->$run_target ($self->_get_dbh, @$args )
887 $_[0]->_get_dbh; # connects or reconnects on pid change, necessary to grab correct txn_depth
888 shift->next::method(@_);
893 Our C<disconnect> method also performs a rollback first if the
894 database is not in C<AutoCommit> mode.
901 # this physical disconnect below might very well throw
902 # in order to unambiguously reset the state - do the cleanup in guard
904 my $g = scope_guard {
907 local $@ if DBIx::Class::_ENV_::UNSTABLE_DOLLARAT;
908 eval { $self->_dbh->disconnect };
912 $self->_dbh_details({});
913 $self->transaction_depth(undef);
914 $self->_dbh_autocommit(undef);
915 $self->savepoints([]);
917 # FIXME - this needs reenabling with the proper "no reset on same DSN" check
918 #$self->_sql_maker(undef); # this may also end up being different
923 $self->_do_connection_actions(disconnect_call_ => $_) for (
924 ( $self->on_disconnect_call || () ),
925 $self->_parse_connect_do ('on_disconnect_do')
928 # stops the "implicit rollback on disconnect" warning
929 $self->_exec_txn_rollback unless $self->_dbh_autocommit;
932 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
933 # collected before leaving this scope. Depending on the code above, this
934 # may very well be just a preventive measure guarding future modifications
938 =head2 with_deferred_fk_checks
942 =item Arguments: C<$coderef>
944 =item Return Value: The return value of $coderef
948 Storage specific method to run the code ref with FK checks deferred or
949 in MySQL's case disabled entirely.
953 # Storage subclasses should override this
954 sub with_deferred_fk_checks {
955 #my ($self, $sub) = @_;
963 =item Arguments: none
965 =item Return Value: 1|0
969 Verifies that the current database handle is active and ready to execute
970 an SQL statement (e.g. the connection did not get stale, server is still
971 answering, etc.) This method is used internally by L</dbh>.
976 return 0 unless $_[0]->_seems_connected;
979 local $_[0]->_dbh->{RaiseError} = 1;
984 sub _seems_connected {
985 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
989 $_[0]->_dbh->FETCH('Active')
993 # explicitly reset all state
999 ($_[0]->_dbh || return 0)->ping;
1002 sub ensure_connected {
1003 $_[0]->connected || ( $_[0]->_populate_dbh && 1 );
1008 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
1009 is guaranteed to be healthy by implicitly calling L</connected>, and if
1010 necessary performing a reconnection before returning. Keep in mind that this
1011 is very B<expensive> on some database engines. Consider using L</dbh_do>
1017 # maybe save a ping call
1019 ? ( $_[0]->ensure_connected and $_[0]->_dbh )
1020 : $_[0]->_populate_dbh
1024 # this is the internal "get dbh or connect (don't check)" method
1026 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
1027 $_[0]->_dbh || $_[0]->_populate_dbh;
1030 # *DELIBERATELY* not a setter (for the time being)
1031 # Too intertwined with everything else for any kind of sanity
1035 $self->throw_exception('sql_maker() is not a setter method') if @_;
1037 unless ($self->_sql_maker) {
1038 my $sql_maker_class = $self->sql_maker_class;
1040 my %opts = %{$self->_sql_maker_opts||{}};
1042 $opts{limit_dialect}
1044 $self->sql_limit_dialect
1047 my $s_class = (ref $self) || $self;
1049 "Your storage class ($s_class) does not set sql_limit_dialect and you "
1050 . 'have not supplied an explicit limit_dialect in your connection_info. '
1051 . 'DBIC will attempt to use the GenericSubQ dialect, which works on most '
1052 . 'databases but can be (and often is) painfully slow. '
1053 . "Please file an RT ticket against '$s_class'"
1054 ) if $self->_dbi_connect_info->[0];
1060 my ($quote_char, $name_sep);
1062 if ($opts{quote_names}) {
1063 $quote_char = (delete $opts{quote_char}) || $self->sql_quote_char || do {
1064 my $s_class = (ref $self) || $self;
1066 "You requested 'quote_names' but your storage class ($s_class) does "
1067 . 'not explicitly define a default sql_quote_char and you have not '
1068 . 'supplied a quote_char as part of your connection_info. DBIC will '
1069 .q{default to the ANSI SQL standard quote '"', which works most of }
1070 . "the time. Please file an RT ticket against '$s_class'."
1076 $name_sep = (delete $opts{name_sep}) || $self->sql_name_sep;
1079 $self->_sql_maker($sql_maker_class->new(
1080 bindtype=>'columns',
1081 array_datatypes => 1,
1082 limit_dialect => $dialect,
1083 ($quote_char ? (quote_char => $quote_char) : ()),
1084 name_sep => ($name_sep || '.'),
1088 return $self->_sql_maker;
1091 # nothing to do by default
1097 # reset internal states
1098 # also in case ->connected failed we might get sent here
1101 $_[0]->_dbh($_[0]->_connect);
1103 $_[0]->_conn_pid($$) unless DBIx::Class::_ENV_::BROKEN_FORK; # on win32 these are in fact threads
1105 $_[0]->_determine_driver;
1107 # Always set the transaction depth on connect, since
1108 # there is no transaction in progress by definition
1109 $_[0]->transaction_depth( $_[0]->_dbh_autocommit ? 0 : 1 );
1111 $_[0]->_run_connection_actions unless $_[0]->{_in_determine_driver};
1116 sub _run_connection_actions {
1118 $_[0]->_do_connection_actions(connect_call_ => $_) for (
1119 ( $_[0]->on_connect_call || () ),
1120 $_[0]->_parse_connect_do ('on_connect_do'),
1126 sub set_use_dbms_capability {
1127 $_[0]->set_inherited ($_[1], $_[2]);
1130 sub get_use_dbms_capability {
1131 my ($self, $capname) = @_;
1133 my $use = $self->get_inherited ($capname);
1136 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1140 sub set_dbms_capability {
1141 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1144 sub get_dbms_capability {
1145 my ($self, $capname) = @_;
1147 my $cap = $self->_dbh_details->{capability}{$capname};
1149 unless (defined $cap) {
1150 if (my $meth = $self->can ("_determine$capname")) {
1151 $cap = $self->$meth ? 1 : 0;
1157 $self->set_dbms_capability ($capname, $cap);
1166 # FIXME - ideally this needs to be an ||= assignment, and the final
1167 # assignment at the end of this do{} should be gone entirely. However
1168 # this confuses CXSA: https://rt.cpan.org/Ticket/Display.html?id=103296
1169 $self->_dbh_details->{info} || do {
1171 # this guarantees that problematic conninfo won't be hidden
1172 # by the try{} below
1173 $self->ensure_connected;
1177 my $server_version = dbic_internal_try {
1178 $self->_get_server_version
1180 # driver determination *may* use this codepath
1181 # in which case we must rethrow
1182 $self->throw_exception($_) if $self->{_in_determine_driver};
1184 # $server_version on failure
1188 if (defined $server_version) {
1189 $info->{dbms_version} = $server_version;
1191 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1192 my @verparts = split (/\./, $numeric_version);
1198 # consider only up to 3 version parts, iff not more than 3 digits
1200 while (@verparts && @use_parts < 3) {
1201 my $p = shift @verparts;
1203 push @use_parts, $p;
1205 push @use_parts, 0 while @use_parts < 3;
1207 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1211 $self->_dbh_details->{info} = $info;
1215 sub _get_server_version {
1216 shift->_dbh_get_info('SQL_DBMS_VER');
1220 my ($self, $info) = @_;
1222 if ($info =~ /[^0-9]/) {
1223 require DBI::Const::GetInfoType;
1224 $info = $DBI::Const::GetInfoType::GetInfoType{$info};
1225 $self->throw_exception("Info type '$_[1]' not provided by DBI::Const::GetInfoType")
1226 unless defined $info;
1229 $self->_get_dbh->get_info($info);
1232 sub _describe_connection {
1233 require DBI::Const::GetInfoReturn;
1239 $drv = $self->_extract_driver_from_connect_info;
1240 $self->ensure_connected;
1243 $drv = "DBD::$drv" if $drv;
1246 DBIC_DSN => $self->_dbi_connect_info->[0],
1247 DBI_VER => DBI->VERSION,
1248 DBIC_VER => DBIx::Class->VERSION,
1249 DBIC_DRIVER => ref $self,
1252 DBD_VER => dbic_internal_try { $drv->VERSION },
1256 # try to grab data even if we never managed to connect
1257 # will cover us in cases of an oddly broken half-connect
1259 #keys %DBI::Const::GetInfoType::GetInfoType,
1261 SQL_CURSOR_COMMIT_BEHAVIOR
1262 SQL_CURSOR_ROLLBACK_BEHAVIOR
1263 SQL_CURSOR_SENSITIVITY
1264 SQL_DATA_SOURCE_NAME
1267 SQL_DEFAULT_TXN_ISOLATION
1272 SQL_EXPRESSIONS_IN_ORDERBY
1275 SQL_IDENTIFIER_QUOTE_CHAR
1276 SQL_MAX_CATALOG_NAME_LEN
1277 SQL_MAX_COLUMN_NAME_LEN
1278 SQL_MAX_IDENTIFIER_LEN
1279 SQL_MAX_TABLE_NAME_LEN
1280 SQL_MULTIPLE_ACTIVE_TXN
1281 SQL_MULT_RESULT_SETS
1282 SQL_NEED_LONG_DATA_LEN
1283 SQL_NON_NULLABLE_COLUMNS
1285 SQL_QUALIFIER_NAME_SEPARATOR
1286 SQL_QUOTED_IDENTIFIER_CASE
1288 SQL_TXN_ISOLATION_OPTION
1291 # some drivers barf on things they do not know about instead
1292 # of returning undef
1293 my $v = dbic_internal_try { $self->_dbh_get_info($inf) };
1294 next unless defined $v;
1296 #my $key = sprintf( '%s(%s)', $inf, $DBI::Const::GetInfoType::GetInfoType{$inf} );
1297 my $expl = DBI::Const::GetInfoReturn::Explain($inf, $v);
1298 $res->{$inf} = DBI::Const::GetInfoReturn::Format($inf, $v) . ( $expl ? " ($expl)" : '' );
1304 sub _determine_driver {
1307 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1308 my $started_connected = 0;
1309 local $self->{_in_determine_driver} = 1;
1311 if (ref($self) eq __PACKAGE__) {
1313 if ($self->_dbh) { # we are connected
1314 $driver = $self->_dbh->{Driver}{Name};
1315 $started_connected = 1;
1318 $driver = $self->_extract_driver_from_connect_info;
1322 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1323 if ($self->load_optional_class($storage_class)) {
1324 mro::set_mro($storage_class, 'c3');
1325 bless $self, $storage_class;
1329 $self->_warn_undetermined_driver(
1330 'This version of DBIC does not yet seem to supply a driver for '
1331 . "your particular RDBMS and/or connection method ('$driver')."
1336 $self->_warn_undetermined_driver(
1337 'Unable to extract a driver name from connect info - this '
1338 . 'should not have happened.'
1343 $self->_driver_determined(1);
1345 Class::C3->reinitialize() if DBIx::Class::_ENV_::OLD_MRO;
1347 if ($self->can('source_bind_attributes')) {
1348 $self->throw_exception(
1349 "Your storage subclass @{[ ref $self ]} provides (or inherits) the method "
1350 . 'source_bind_attributes() for which support has been removed as of Jan 2013. '
1351 . 'If you are not sure how to proceed please contact the development team via '
1352 . DBIx::Class::_ENV_::HELP_URL
1356 $self->_init; # run driver-specific initializations
1358 $self->_run_connection_actions
1359 if !$started_connected && defined $self->_dbh;
1363 sub _extract_driver_from_connect_info {
1368 # if connect_info is a CODEREF, we have no choice but to connect
1370 ref $self->_dbi_connect_info->[0]
1372 reftype $self->_dbi_connect_info->[0] eq 'CODE'
1374 $self->_populate_dbh;
1375 $drv = $self->_dbh->{Driver}{Name};
1378 # try to use dsn to not require being connected, the driver may still
1379 # force a connection later in _rebless to determine version
1380 # (dsn may not be supplied at all if all we do is make a mock-schema)
1381 ($drv) = ($self->_dbi_connect_info->[0] || '') =~ /^dbi:([^:]+):/i;
1382 $drv ||= $ENV{DBI_DRIVER};
1388 sub _determine_connector_driver {
1389 my ($self, $conn) = @_;
1391 my $dbtype = $self->_get_rdbms_name;
1394 $self->_warn_undetermined_driver(
1395 'Unable to retrieve RDBMS type (SQL_DBMS_NAME) of the engine behind your '
1396 . "$conn connector - this should not have happened."
1401 $dbtype =~ s/\W/_/gi;
1403 my $subclass = "DBIx::Class::Storage::DBI::${conn}::${dbtype}";
1404 return if $self->isa($subclass);
1406 if ($self->load_optional_class($subclass)) {
1407 bless $self, $subclass;
1411 $self->_warn_undetermined_driver(
1412 'This version of DBIC does not yet seem to supply a driver for '
1413 . "your particular RDBMS and/or connection method ('$conn/$dbtype')."
1418 sub _get_rdbms_name { shift->_dbh_get_info('SQL_DBMS_NAME') }
1420 sub _warn_undetermined_driver {
1421 my ($self, $msg) = @_;
1423 require Data::Dumper::Concise;
1425 carp_once ($msg . ' While we will attempt to continue anyway, the results '
1426 . 'are likely to be underwhelming. Please upgrade DBIC, and if this message '
1427 . "does not go away, file a bugreport including the following info:\n"
1428 . Data::Dumper::Concise::Dumper($self->_describe_connection)
1432 sub _do_connection_actions {
1433 my ($self, $method_prefix, $call, @args) = @_;
1436 if (not ref($call)) {
1437 my $method = $method_prefix . $call;
1438 $self->$method(@args);
1440 elsif (ref($call) eq 'CODE') {
1441 $self->$call(@args);
1443 elsif (ref($call) eq 'ARRAY') {
1444 if (ref($call->[0]) ne 'ARRAY') {
1445 $self->_do_connection_actions($method_prefix, $_) for @$call;
1448 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1452 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1456 if ( $method_prefix =~ /^connect/ ) {
1457 # this is an on_connect cycle - we can't just throw while leaving
1458 # a handle in an undefined state in our storage object
1459 # kill it with fire and rethrow
1461 $self->disconnect; # the $dbh is gone, but we still need to reset the rest
1462 $self->throw_exception( $_[0] );
1465 carp "Disconnect action failed: $_[0]";
1472 sub connect_call_do_sql {
1474 $self->_do_query(@_);
1477 sub disconnect_call_do_sql {
1479 $self->_do_query(@_);
1482 =head2 connect_call_datetime_setup
1484 A no-op stub method, provided so that one can always safely supply the
1485 L<connection option|/DBIx::Class specific connection attributes>
1487 on_connect_call => 'datetime_setup'
1489 This way one does not need to know in advance whether the underlying
1490 storage requires any sort of hand-holding when dealing with calendar
1495 sub connect_call_datetime_setup { 1 }
1498 my ($self, $action) = @_;
1500 if (ref $action eq 'CODE') {
1501 $action = $action->($self);
1502 $self->_do_query($_) foreach @$action;
1505 # Most debuggers expect ($sql, @bind), so we need to exclude
1506 # the attribute hash which is the second argument to $dbh->do
1507 # furthermore the bind values are usually to be presented
1508 # as named arrayref pairs, so wrap those here too
1509 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1510 my $sql = shift @do_args;
1511 my $attrs = shift @do_args;
1512 my @bind = map { [ undef, $_ ] } @do_args;
1515 $_[0]->_query_start($sql, \@bind);
1516 $_[1]->do($sql, $attrs, @do_args);
1517 $_[0]->_query_end($sql, \@bind);
1527 my $info = $self->_dbi_connect_info;
1529 $self->throw_exception("You did not provide any connection_info")
1530 unless defined $info->[0];
1532 my ($old_connect_via, $dbh);
1534 local $DBI::connect_via = 'connect' if $INC{'Apache/DBI.pm'} && $ENV{MOD_PERL};
1536 # this odd anonymous coderef dereference is in fact really
1537 # necessary to avoid the unwanted effect described in perl5
1540 # in addition the coderef itself can't reside inside the try{} block below
1541 # as it somehow triggers a leak under perl -d
1542 my $dbh_error_handler_installer = sub {
1543 weaken (my $weak_self = $_[0]);
1545 # the coderef is blessed so we can distinguish it from externally
1546 # supplied handles (which must be preserved)
1547 $_[1]->{HandleError} = bless sub {
1549 $weak_self->throw_exception("DBI Exception: $_[0]");
1552 # the handler may be invoked by something totally out of
1554 DBIx::Class::Exception->throw("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1556 }, '__DBIC__DBH__ERROR__HANDLER__';
1560 if(ref $info->[0] eq 'CODE') {
1561 $dbh = $info->[0]->();
1565 $dbh = DBI->connect(@$info);
1568 die $DBI::errstr unless $dbh;
1570 die sprintf ("%s fresh DBI handle with a *false* 'Active' attribute. "
1571 . 'This handle is disconnected as far as DBIC is concerned, and we can '
1573 ref $info->[0] eq 'CODE'
1574 ? "Connection coderef $info->[0] returned a"
1575 : 'DBI->connect($schema->storage->connect_info) resulted in a'
1576 ) unless $dbh->FETCH('Active');
1578 # sanity checks unless asked otherwise
1579 unless ($self->unsafe) {
1581 $self->throw_exception(
1582 'Refusing clobbering of {HandleError} installed on externally supplied '
1583 ."DBI handle $dbh. Either remove the handler or use the 'unsafe' attribute."
1584 ) if $dbh->{HandleError} and ref $dbh->{HandleError} ne '__DBIC__DBH__ERROR__HANDLER__';
1586 # Default via _default_dbi_connect_attributes is 1, hence it was an explicit
1587 # request, or an external handle. Complain and set anyway
1588 unless ($dbh->{RaiseError}) {
1589 carp( ref $info->[0] eq 'CODE'
1591 ? "The 'RaiseError' of the externally supplied DBI handle is set to false. "
1592 ."DBIx::Class will toggle it back to true, unless the 'unsafe' connect "
1593 .'attribute has been supplied'
1595 : 'RaiseError => 0 supplied in your connection_info, without an explicit '
1596 .'unsafe => 1. Toggling RaiseError back to true'
1599 $dbh->{RaiseError} = 1;
1602 $dbh_error_handler_installer->($self, $dbh);
1606 $self->throw_exception("DBI Connection failed: $_")
1609 $self->_dbh_autocommit($dbh->{AutoCommit});
1614 # this means we have not yet connected and do not know the AC status
1615 # (e.g. coderef $dbh), need a full-fledged connection check
1616 if (! defined $_[0]->_dbh_autocommit) {
1617 $_[0]->ensure_connected;
1619 # Otherwise simply connect or re-connect on pid changes
1624 shift->next::method(@_);
1627 sub _exec_txn_begin {
1630 # if the user is utilizing txn_do - good for him, otherwise we need to
1631 # ensure that the $dbh is healthy on BEGIN.
1632 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1633 # will be replaced by a failure of begin_work itself (which will be
1634 # then retried on reconnect)
1635 if ($self->{_in_do_block}) {
1636 $self->_dbh->begin_work;
1638 $self->dbh_do(sub { $_[1]->begin_work });
1645 $self->throw_exception("Unable to txn_commit() on a disconnected storage")
1646 unless $self->_seems_connected;
1648 # esoteric case for folks using external $dbh handles
1649 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1650 carp "Storage transaction_depth 0 does not match "
1651 ."false AutoCommit of $self->{_dbh}, attempting COMMIT anyway";
1652 $self->transaction_depth(1);
1655 $self->next::method(@_);
1657 # if AutoCommit is disabled txn_depth never goes to 0
1658 # as a new txn is started immediately on commit
1659 $self->transaction_depth(1) if (
1660 !$self->transaction_depth
1662 defined $self->_dbh_autocommit
1664 ! $self->_dbh_autocommit
1668 sub _exec_txn_commit {
1669 shift->_dbh->commit;
1675 # do a minimal connectivity check due to weird shit like
1676 # https://rt.cpan.org/Public/Bug/Display.html?id=62370
1677 $self->throw_exception("lost connection to storage")
1678 unless $self->_seems_connected;
1680 # esoteric case for folks using external $dbh handles
1681 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1682 carp "Storage transaction_depth 0 does not match "
1683 ."false AutoCommit of $self->{_dbh}, attempting ROLLBACK anyway";
1684 $self->transaction_depth(1);
1687 $self->next::method(@_);
1689 # if AutoCommit is disabled txn_depth never goes to 0
1690 # as a new txn is started immediately on commit
1691 $self->transaction_depth(1) if (
1692 !$self->transaction_depth
1694 defined $self->_dbh_autocommit
1696 ! $self->_dbh_autocommit
1700 sub _exec_txn_rollback {
1701 shift->_dbh->rollback;
1704 # generate the DBI-specific stubs, which then fallback to ::Storage proper
1705 quote_sub __PACKAGE__ . "::$_" => sprintf (<<'EOS', $_) for qw(svp_begin svp_release svp_rollback);
1706 $_[0]->throw_exception('Unable to %s() on a disconnected storage')
1707 unless $_[0]->_seems_connected;
1708 shift->next::method(@_);
1711 # This used to be the top-half of _execute. It was split out to make it
1712 # easier to override in NoBindVars without duping the rest. It takes up
1713 # all of _execute's args, and emits $sql, @bind.
1714 sub _prep_for_execute {
1715 #my ($self, $op, $ident, $args) = @_;
1716 return shift->_gen_sql_bind(@_)
1720 my ($self, $op, $ident, $args) = @_;
1722 my ($colinfos, $from);
1723 if ( blessed($ident) ) {
1724 $from = $ident->from;
1725 $colinfos = $ident->columns_info;
1729 ($sql, @$bind) = $self->sql_maker->$op( ($from || $ident), @$args );
1731 $bind = $self->_resolve_bindattrs(
1732 $ident, [ @{$args->[2]{bind}||[]}, @$bind ], $colinfos
1736 ! $ENV{DBIC_DT_SEARCH_OK}
1745 $_->[1]->isa('DateTime')
1748 carp_unique 'DateTime objects passed to search() are not supported '
1749 . 'properly (InflateColumn::DateTime formats and settings are not '
1750 . 'respected.) See ".. format a DateTime object for searching?" in '
1751 . 'DBIx::Class::Manual::FAQ. To disable this warning for good '
1752 . 'set $ENV{DBIC_DT_SEARCH_OK} to true'
1755 return( $sql, $bind );
1758 sub _resolve_bindattrs {
1759 my ($self, $ident, $bind, $colinfos) = @_;
1761 my $resolve_bindinfo = sub {
1762 #my $infohash = shift;
1764 $colinfos ||= { %{ $self->_resolve_column_info($ident) } };
1767 if (my $col = $_[0]->{dbic_colname}) {
1768 $ret = { %{$_[0]} };
1770 $ret->{sqlt_datatype} ||= $colinfos->{$col}{data_type}
1771 if $colinfos->{$col}{data_type};
1773 $ret->{sqlt_size} ||= $colinfos->{$col}{size}
1774 if $colinfos->{$col}{size};
1781 ( ref $_ ne 'ARRAY' or @$_ != 2 ) ? [ {}, $_ ]
1782 : ( ! defined $_->[0] ) ? [ {}, $_->[1] ]
1783 : (ref $_->[0] eq 'HASH') ? [(
1786 exists $_->[0]{dbd_attrs}
1788 $_->[0]{sqlt_datatype}
1790 : $resolve_bindinfo->($_->[0])
1793 : (ref $_->[0] eq 'SCALAR') ? [ { sqlt_datatype => ${$_->[0]} }, $_->[1] ]
1794 : [ $resolve_bindinfo->(
1795 { dbic_colname => $_->[0] }
1800 sub _format_for_trace {
1801 #my ($self, $bind) = @_;
1803 ### Turn @bind from something like this:
1804 ### ( [ "artist", 1 ], [ \%attrs, 3 ] )
1806 ### ( "'1'", "'3'" )
1809 defined( $_ && $_->[1] )
1816 my ( $self, $sql, $bind ) = @_;
1818 $self->debugobj->query_start( $sql, $self->_format_for_trace($bind) )
1823 my ( $self, $sql, $bind ) = @_;
1825 $self->debugobj->query_end( $sql, $self->_format_for_trace($bind) )
1829 sub _dbi_attrs_for_bind {
1830 #my ($self, $ident, $bind) = @_;
1834 exists $_->{dbd_attrs} ? $_->{dbd_attrs}
1836 : ! $_->{sqlt_datatype} ? undef
1840 # cache the result in the dbh_details hash, as it (usually) can not change
1841 # unless we connect to something else
1842 # FIXME: for the time being Oracle is an exception, pending a rewrite of
1844 my $cache = $_[0]->_dbh_details->{_datatype_map_cache} ||= {};
1846 $cache->{$_->{sqlt_datatype}} = $_[0]->bind_attribute_by_data_type($_->{sqlt_datatype})
1847 if ! exists $cache->{$_->{sqlt_datatype}};
1849 $cache->{$_->{sqlt_datatype}};
1851 } } map { $_->[0] } @{$_[2]} ];
1855 my ($self, $op, $ident, @args) = @_;
1857 my ($sql, $bind) = $self->_prep_for_execute($op, $ident, \@args);
1859 # not even a PID check - we do not care about the state of the _dbh.
1860 # All we need is to get the appropriate drivers loaded if they aren't
1861 # already so that the assumption in ad7c50fc26e holds
1862 $self->_populate_dbh unless $self->_dbh;
1864 $self->dbh_do( _dbh_execute => # retry over disconnects
1867 $self->_dbi_attrs_for_bind($ident, $bind),
1872 my ($self, $dbh, $sql, $bind, $bind_attrs) = @_;
1874 $self->_query_start( $sql, $bind );
1876 my $sth = $self->_bind_sth_params(
1877 $self->_prepare_sth($dbh, $sql),
1882 # Can this fail without throwing an exception anyways???
1883 my $rv = $sth->execute();
1884 $self->throw_exception(
1885 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1888 $self->_query_end( $sql, $bind );
1890 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1894 my ($self, $dbh, $sql) = @_;
1896 # 3 is the if_active parameter which avoids active sth re-use
1897 my $sth = $self->disable_sth_caching
1898 ? $dbh->prepare($sql)
1899 : $dbh->prepare_cached($sql, {}, 3);
1901 # XXX You would think RaiseError would make this impossible,
1902 # but apparently that's not true :(
1903 $self->throw_exception(
1906 sprintf( "\$dbh->prepare() of '%s' through %s failed *silently* without "
1907 .'an exception and/or setting $dbh->errstr',
1909 ? substr($sql, 0, 20) . '...'
1912 'DBD::' . $dbh->{Driver}{Name},
1919 sub _bind_sth_params {
1920 my ($self, $sth, $bind, $bind_attrs) = @_;
1922 for my $i (0 .. $#$bind) {
1923 if (ref $bind->[$i][1] eq 'SCALAR') { # any scalarrefs are assumed to be bind_inouts
1924 $sth->bind_param_inout(
1925 $i + 1, # bind params counts are 1-based
1927 $bind->[$i][0]{dbd_size} || $self->_max_column_bytesize($bind->[$i][0]), # size
1932 # FIXME SUBOPTIMAL - DBI needs fixing to always stringify regardless of DBD
1933 my $v = ( length ref $bind->[$i][1] and is_plain_value $bind->[$i][1] )
1940 # The temp-var is CRUCIAL - DO NOT REMOVE IT, breaks older DBD::SQLite RT#79576
1950 sub _prefetch_autovalues {
1951 my ($self, $source, $colinfo, $to_insert) = @_;
1954 for my $col (keys %$colinfo) {
1956 $colinfo->{$col}{auto_nextval}
1959 ! exists $to_insert->{$col}
1961 is_literal_value($to_insert->{$col})
1964 $values{$col} = $self->_sequence_fetch(
1966 ( $colinfo->{$col}{sequence} ||=
1967 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1977 my ($self, $source, $to_insert) = @_;
1979 my $col_infos = $source->columns_info;
1981 my $prefetched_values = $self->_prefetch_autovalues($source, $col_infos, $to_insert);
1983 # fuse the values, but keep a separate list of prefetched_values so that
1984 # they can be fused once again with the final return
1985 $to_insert = { %$to_insert, %$prefetched_values };
1987 # FIXME - we seem to assume undef values as non-supplied. This is wrong.
1988 # Investigate what does it take to s/defined/exists/
1989 my %pcols = map { $_ => 1 } $source->primary_columns;
1990 my (%retrieve_cols, $autoinc_supplied, $retrieve_autoinc_col);
1991 for my $col ($source->columns) {
1992 if ($col_infos->{$col}{is_auto_increment}) {
1993 $autoinc_supplied ||= 1 if defined $to_insert->{$col};
1994 $retrieve_autoinc_col ||= $col unless $autoinc_supplied;
1997 # nothing to retrieve when explicit values are supplied
1999 defined $to_insert->{$col} and ! is_literal_value($to_insert->{$col})
2002 # the 'scalar keys' is a trick to preserve the ->columns declaration order
2003 $retrieve_cols{$col} = scalar keys %retrieve_cols if (
2006 $col_infos->{$col}{retrieve_on_insert}
2010 local $self->{_autoinc_supplied_for_op} = $autoinc_supplied;
2011 local $self->{_perform_autoinc_retrieval} = $retrieve_autoinc_col;
2013 my ($sqla_opts, @ir_container);
2014 if (%retrieve_cols and $self->_use_insert_returning) {
2015 $sqla_opts->{returning_container} = \@ir_container
2016 if $self->_use_insert_returning_bound;
2018 $sqla_opts->{returning} = [
2019 sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols
2023 my ($rv, $sth) = $self->_execute('insert', $source, $to_insert, $sqla_opts);
2025 my %returned_cols = %$to_insert;
2026 if (my $retlist = $sqla_opts->{returning}) { # if IR is supported - we will get everything in one set
2028 unless( @ir_container ) {
2031 # FIXME - need to investigate why Caelum silenced this in 4d4dc518
2032 local $SIG{__WARN__} = sub {};
2034 @ir_container = $sth->fetchrow_array;
2038 # Evict the $sth from the cache in case we got here, since the finish()
2039 # is crucial, at least on older Firebirds, possibly on other engines too
2041 # It would be too complex to make this a proper subclass override,
2042 # and besides we already take the try{} penalty, adding a catch that
2043 # triggers infrequently is a no-brainer
2045 if( my $kids = $self->_dbh->{CachedKids} ) {
2046 $kids->{$_} == $sth and delete $kids->{$_}
2052 @returned_cols{@$retlist} = @ir_container if @ir_container;
2055 # pull in PK if needed and then everything else
2056 if (my @missing_pri = grep { $pcols{$_} } keys %retrieve_cols) {
2058 $self->throw_exception( "Missing primary key but Storage doesn't support last_insert_id" )
2059 unless $self->can('last_insert_id');
2061 my @pri_values = $self->last_insert_id($source, @missing_pri);
2063 $self->throw_exception( "Can't get last insert id" )
2064 unless (@pri_values == @missing_pri);
2066 @returned_cols{@missing_pri} = @pri_values;
2067 delete @retrieve_cols{@missing_pri};
2070 # if there is more left to pull
2071 if (%retrieve_cols) {
2072 $self->throw_exception(
2073 'Unable to retrieve additional columns without a Primary Key on ' . $source->source_name
2076 my @left_to_fetch = sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols;
2078 my $cur = DBIx::Class::ResultSet->new($source, {
2079 where => { map { $_ => $returned_cols{$_} } (keys %pcols) },
2080 select => \@left_to_fetch,
2083 @returned_cols{@left_to_fetch} = $cur->next;
2085 $self->throw_exception('Duplicate row returned for PK-search after fresh insert')
2086 if scalar $cur->next;
2090 return { %$prefetched_values, %returned_cols };
2095 'insert_bulk() should have never been exposed as a public method and '
2096 . 'calling it is depecated as of Aug 2014. If you believe having a genuine '
2097 . 'use for this method please contact the development team via '
2098 . DBIx::Class::_ENV_::HELP_URL
2101 return '0E0' unless @{$_[3]||[]};
2103 shift->_insert_bulk(@_);
2107 my ($self, $source, $cols, $data) = @_;
2109 $self->throw_exception('Calling _insert_bulk without a dataset to process makes no sense')
2110 unless @{$data||[]};
2112 my $colinfos = $source->columns_info($cols);
2114 local $self->{_autoinc_supplied_for_op} =
2115 (grep { $_->{is_auto_increment} } values %$colinfos)
2120 # get a slice type index based on first row of data
2121 # a "column" in this context may refer to more than one bind value
2122 # e.g. \[ '?, ?', [...], [...] ]
2124 # construct the value type index - a description of values types for every
2125 # per-column slice of $data:
2127 # nonexistent - nonbind literal
2129 # [] of bindattrs - resolved attribute(s) of bind(s) passed via literal+bind \[] combo
2131 # also construct the column hash to pass to the SQL generator. For plain
2132 # (non literal) values - convert the members of the first row into a
2133 # literal+bind combo, with extra positional info in the bind attr hashref.
2134 # This will allow us to match the order properly, and is so contrived
2135 # because a user-supplied literal/bind (or something else specific to a
2136 # resultsource and/or storage driver) can inject extra binds along the
2137 # way, so one can't rely on "shift positions" ordering at all. Also we
2138 # can't just hand SQLA a set of some known "values" (e.g. hashrefs that
2139 # can be later matched up by address), because we want to supply a real
2140 # value on which perhaps e.g. datatype checks will be performed
2141 my ($proto_data, $serialized_bind_type_by_col_idx);
2142 for my $col_idx (0..$#$cols) {
2143 my $colname = $cols->[$col_idx];
2144 if (ref $data->[0][$col_idx] eq 'SCALAR') {
2145 # no bind value at all - no type
2147 $proto_data->{$colname} = $data->[0][$col_idx];
2149 elsif (ref $data->[0][$col_idx] eq 'REF' and ref ${$data->[0][$col_idx]} eq 'ARRAY' ) {
2150 # repack, so we don't end up mangling the original \[]
2151 my ($sql, @bind) = @${$data->[0][$col_idx]};
2153 # normalization of user supplied stuff
2154 my $resolved_bind = $self->_resolve_bindattrs(
2155 $source, \@bind, $colinfos,
2158 # store value-less (attrs only) bind info - we will be comparing all
2159 # supplied binds against this for sanity
2160 $serialized_bind_type_by_col_idx->{$col_idx} = serialize [ map { $_->[0] } @$resolved_bind ];
2162 $proto_data->{$colname} = \[ $sql, map { [
2163 # inject slice order to use for $proto_bind construction
2164 { %{$resolved_bind->[$_][0]}, _bind_data_slice_idx => $col_idx, _literal_bind_subindex => $_+1 }
2166 $resolved_bind->[$_][1]
2171 $serialized_bind_type_by_col_idx->{$col_idx} = undef;
2173 $proto_data->{$colname} = \[ '?', [
2174 { dbic_colname => $colname, _bind_data_slice_idx => $col_idx }
2176 $data->[0][$col_idx]
2181 my ($sql, $proto_bind) = $self->_prep_for_execute (
2187 if (! @$proto_bind and keys %$serialized_bind_type_by_col_idx) {
2188 # if the bindlist is empty and we had some dynamic binds, this means the
2189 # storage ate them away (e.g. the NoBindVars component) and interpolated
2190 # them directly into the SQL. This obviously can't be good for multi-inserts
2191 $self->throw_exception('Unable to invoke fast-path insert without storage placeholder support');
2195 # FIXME - devise a flag "no babysitting" or somesuch to shut this off
2197 # use an error reporting closure for convenience (less to pass)
2198 my $bad_slice_report_cref = sub {
2199 my ($msg, $r_idx, $c_idx) = @_;
2200 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
2204 require Data::Dumper::Concise;
2205 local $Data::Dumper::Maxdepth = 5;
2206 Data::Dumper::Concise::Dumper ({
2207 map { $cols->[$_] =>
2215 for my $col_idx (0..$#$cols) {
2216 my $reference_val = $data->[0][$col_idx];
2218 for my $row_idx (1..$#$data) { # we are comparing against what we got from [0] above, hence start from 1
2219 my $val = $data->[$row_idx][$col_idx];
2221 if (! exists $serialized_bind_type_by_col_idx->{$col_idx}) { # literal no binds
2222 if (ref $val ne 'SCALAR') {
2223 $bad_slice_report_cref->(
2224 "Incorrect value (expecting SCALAR-ref \\'$$reference_val')",
2229 elsif ($$val ne $$reference_val) {
2230 $bad_slice_report_cref->(
2231 "Inconsistent literal SQL value (expecting \\'$$reference_val')",
2237 elsif (! defined $serialized_bind_type_by_col_idx->{$col_idx} ) { # regular non-literal value
2238 if (is_literal_value($val)) {
2239 $bad_slice_report_cref->("Literal SQL found where a plain bind value is expected", $row_idx, $col_idx);
2242 else { # binds from a \[], compare type and attrs
2243 if (ref $val ne 'REF' or ref $$val ne 'ARRAY') {
2244 $bad_slice_report_cref->(
2245 "Incorrect value (expecting ARRAYREF-ref \\['${$reference_val}->[0]', ... ])",
2250 # start drilling down and bail out early on identical refs
2252 $reference_val != $val
2254 $$reference_val != $$val
2256 if (${$val}->[0] ne ${$reference_val}->[0]) {
2257 $bad_slice_report_cref->(
2258 "Inconsistent literal/bind SQL (expecting \\['${$reference_val}->[0]', ... ])",
2263 # need to check the bind attrs - a bind will happen only once for
2264 # the entire dataset, so any changes further down will be ignored.
2266 $serialized_bind_type_by_col_idx->{$col_idx}
2271 @{$self->_resolve_bindattrs(
2272 $source, [ @{$$val}[1 .. $#$$val] ], $colinfos,
2276 $bad_slice_report_cref->(
2277 'Differing bind attributes on literal/bind values not supported',
2287 # neither _dbh_execute_for_fetch, nor _dbh_execute_inserts_with_no_binds
2288 # are atomic (even if execute_for_fetch is a single call). Thus a safety
2290 my $guard = $self->txn_scope_guard;
2292 $self->_query_start( $sql, @$proto_bind ? [[ {} => '__BULK_INSERT__' ]] : () );
2293 my $sth = $self->_prepare_sth($self->_dbh, $sql);
2296 # proto bind contains the information on which pieces of $data to pull
2297 # $cols is passed in only for prettier error-reporting
2298 $self->_dbh_execute_for_fetch( $source, $sth, $proto_bind, $cols, $data );
2301 # bind_param_array doesn't work if there are no binds
2302 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
2306 $self->_query_end( $sql, @$proto_bind ? [[ {} => '__BULK_INSERT__' ]] : () );
2310 return wantarray ? ($rv, $sth, @$proto_bind) : $rv;
2313 # execute_for_fetch is capable of returning data just fine (it means it
2314 # can be used for INSERT...RETURNING and UPDATE...RETURNING. Since this
2315 # is the void-populate fast-path we will just ignore this altogether
2316 # for the time being.
2317 sub _dbh_execute_for_fetch {
2318 my ($self, $source, $sth, $proto_bind, $cols, $data) = @_;
2320 # If we have any bind attributes to take care of, we will bind the
2321 # proto-bind data (which will never be used by execute_for_fetch)
2322 # However since column bindtypes are "sticky", this is sufficient
2323 # to get the DBD to apply the bindtype to all values later on
2324 my $bind_attrs = $self->_dbi_attrs_for_bind($source, $proto_bind);
2326 for my $i (0 .. $#$proto_bind) {
2328 $i+1, # DBI bind indexes are 1-based
2329 $proto_bind->[$i][1],
2331 ) if defined $bind_attrs->[$i];
2334 # At this point $data slots named in the _bind_data_slice_idx of
2335 # each piece of $proto_bind are either \[]s or plain values to be
2336 # passed in. Construct the dispensing coderef. *NOTE* the order
2337 # of $data will differ from this of the ?s in the SQL (due to
2338 # alphabetical ordering by colname). We actually do want to
2339 # preserve this behavior so that prepare_cached has a better
2340 # chance of matching on unrelated calls
2342 my $fetch_row_idx = -1; # saner loop this way
2343 my $fetch_tuple = sub {
2344 return undef if ++$fetch_row_idx > $#$data;
2347 my $v = ! defined $_->{_literal_bind_subindex}
2349 ? $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]
2351 # There are no attributes to resolve here - we already did everything
2352 # when we constructed proto_bind. However we still want to sanity-check
2353 # what the user supplied, so pass stuff through to the resolver *anyway*
2354 : $self->_resolve_bindattrs (
2355 undef, # a fake rsrc
2356 [ ${ $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]}->[ $_->{_literal_bind_subindex} ] ],
2357 {}, # a fake column_info bag
2361 # FIXME SUBOPTIMAL - DBI needs fixing to always stringify regardless of DBD
2362 # For the time being forcibly stringify whatever is stringifiable
2365 ( !length ref $v or ! ($vref = is_plain_value $v) ) ? $v
2366 : defined blessed( $$vref ) ? "$$vref"
2369 } map { $_->[0] } @$proto_bind ];
2372 my $tuple_status = [];
2375 $rv = $sth->execute_for_fetch(
2384 # Not all DBDs are create equal. Some throw on error, some return
2385 # an undef $rv, and some set $sth->err - try whatever we can
2386 $err = ($sth->errstr || 'UNKNOWN ERROR ($sth->errstr is unset)') if (
2389 ( !defined $rv or $sth->err )
2392 # Statement must finish even if there was an exception.
2397 $err = shift unless defined $err
2402 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
2404 $self->throw_exception("Unexpected populate error: $err")
2405 if ($i > $#$tuple_status);
2407 require Data::Dumper::Concise;
2408 $self->throw_exception(sprintf "execute_for_fetch() aborted with '%s' at populate slice:\n%s",
2409 ($tuple_status->[$i][1] || $err),
2410 Data::Dumper::Concise::Dumper( { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) } ),
2417 sub _dbh_execute_inserts_with_no_binds {
2418 my ($self, $sth, $count) = @_;
2422 my $dbh = $self->_get_dbh;
2423 local $dbh->{RaiseError} = 1;
2424 local $dbh->{PrintError} = 0;
2426 $sth->execute foreach 1..$count;
2432 # Make sure statement is finished even if there was an exception.
2437 $err = shift unless defined $err;
2440 $self->throw_exception($err) if defined $err;
2446 #my ($self, $source, @args) = @_;
2447 shift->_execute('update', @_);
2452 #my ($self, $source, @args) = @_;
2453 shift->_execute('delete', @_);
2458 $self->_execute($self->_select_args(@_));
2461 sub _select_args_to_query {
2464 $self->throw_exception(
2465 "Unable to generate limited query representation with 'software_limit' enabled"
2466 ) if ($_[3]->{software_limit} and ($_[3]->{offset} or $_[3]->{rows}) );
2468 # my ($op, $ident, $select, $cond, $rs_attrs, $rows, $offset)
2469 # = $self->_select_args($ident, $select, $cond, $attrs);
2470 my ($op, $ident, @args) =
2471 $self->_select_args(@_);
2473 # my ($sql, $prepared_bind) = $self->_gen_sql_bind($op, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
2474 my ($sql, $bind) = $self->_gen_sql_bind($op, $ident, \@args);
2476 # reuse the bind arrayref
2477 unshift @{$bind}, "($sql)";
2482 my ($self, $ident, $select, $where, $orig_attrs) = @_;
2484 # FIXME - that kind of caching would be nice to have
2485 # however currently we *may* pass the same $orig_attrs
2486 # with different ident/select/where
2487 # the whole interface needs to be rethought, since it
2488 # was centered around the flawed SQLA API. We can do
2489 # soooooo much better now. But that is also another
2492 # 'select', $orig_attrs->{!args_as_stored_at_the_end_of_this_method!}
2493 #) if $orig_attrs->{!args_as_stored_at_the_end_of_this_method!};
2495 my $sql_maker = $self->sql_maker;
2504 # MySQL actually recommends this approach. I cringe.
2505 $attrs->{rows} ||= $sql_maker->__max_int
2506 if $attrs->{offset};
2508 # see if we will need to tear the prefetch apart to satisfy group_by == select
2509 # this is *extremely tricky* to get right, I am still not sure I did
2511 my ($prefetch_needs_subquery, @limit_args);
2513 if ( $attrs->{_grouped_by_distinct} and $attrs->{collapse} ) {
2514 # we already know there is a valid group_by (we made it) and we know it is
2515 # intended to be based *only* on non-multi stuff
2516 # short circuit the group_by parsing below
2517 $prefetch_needs_subquery = 1;
2520 # The rationale is that even if we do *not* have collapse, we still
2521 # need to wrap the core grouped select/group_by in a subquery
2522 # so that databases that care about group_by/select equivalence
2523 # are happy (this includes MySQL in strict_mode)
2524 # If any of the other joined tables are referenced in the group_by
2525 # however - the user is on their own
2526 ( $prefetch_needs_subquery or ! $attrs->{_simple_passthrough_construction} )
2530 @{$attrs->{group_by}}
2532 my $grp_aliases = dbic_internal_try { # internal_try{} because $attrs->{from} may be unreadable
2533 $self->_resolve_aliastypes_from_select_args({ from => $attrs->{from}, group_by => $attrs->{group_by} })
2536 # no aliases other than our own in group_by
2537 # if there are - do not allow subquery even if limit is present
2538 $prefetch_needs_subquery = ! scalar grep { $_ ne $attrs->{alias} } keys %{ $grp_aliases->{grouping} || {} };
2540 elsif ( $attrs->{rows} && $attrs->{collapse} ) {
2541 # active collapse with a limit - that one is a no-brainer unless
2542 # overruled by a group_by above
2543 $prefetch_needs_subquery = 1;
2546 if ($prefetch_needs_subquery) {
2547 $attrs = $self->_adjust_select_args_for_complex_prefetch ($attrs);
2549 elsif (! $attrs->{software_limit} ) {
2551 $attrs->{rows} || (),
2552 $attrs->{offset} || (),
2556 # try to simplify the joinmap further (prune unreferenced type-single joins)
2558 ! $prefetch_needs_subquery # already pruned
2562 reftype $attrs->{from} eq 'ARRAY'
2564 @{$attrs->{from}} != 1
2566 ($attrs->{from}, $attrs->{_aliastypes}) = $self->_prune_unused_joins ($attrs);
2569 # FIXME this is a gross, inefficient, largely incorrect and fragile hack
2570 # during the result inflation stage we *need* to know what was the aliastype
2571 # map as sqla saw it when the final pieces of SQL were being assembled
2572 # Originally we simply carried around the entirety of $attrs, but this
2573 # resulted in resultsets that are being reused growing continuously, as
2574 # the hash in question grew deeper and deeper.
2575 # Instead hand-pick what to take with us here (we actually don't need much
2576 # at this point just the map itself)
2577 $orig_attrs->{_last_sqlmaker_alias_map} = $attrs->{_aliastypes};
2580 # my $alias2source = $self->_resolve_ident_sources ($ident);
2582 # This would be the point to deflate anything found in $attrs->{where}
2583 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2584 # expect a result object. And all we have is a resultsource (it is trivial
2585 # to extract deflator coderefs via $alias2source above).
2587 # I don't see a way forward other than changing the way deflators are
2588 # invoked, and that's just bad...
2591 return ( 'select', @{$attrs}{qw(from select where)}, $attrs, @limit_args );
2594 # Returns a counting SELECT for a simple count
2595 # query. Abstracted so that a storage could override
2596 # this to { count => 'firstcol' } or whatever makes
2597 # sense as a performance optimization
2599 #my ($self, $source, $rs_attrs) = @_;
2600 return { count => '*' };
2607 =item Arguments: $ident, $select, $condition, $attrs
2611 Handle a SQL select statement.
2617 my ($ident, $select, $condition, $attrs) = @_;
2618 return $self->cursor_class->new($self, \@_, $attrs);
2623 my ($rv, $sth, @bind) = $self->_select(@_);
2624 my @row = $sth->fetchrow_array;
2625 my @nextrow = $sth->fetchrow_array if @row;
2626 if(@row && @nextrow) {
2627 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2629 # Need to call finish() to work round broken DBDs
2634 =head2 sql_limit_dialect
2636 This is an accessor for the default SQL limit dialect used by a particular
2637 storage driver. Can be overridden by supplying an explicit L</limit_dialect>
2638 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2639 see L<DBIx::Class::SQLMaker::LimitDialects>.
2643 sub _dbh_columns_info_for {
2644 my ($self, $dbh, $table) = @_;
2648 if (! DBIx::Class::_ENV_::STRESSTEST_COLUMN_INFO_UNAWARE_STORAGE and $dbh->can('column_info')) {
2650 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2651 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2653 while ( my $info = $sth->fetchrow_hashref() ){
2655 $column_info{data_type} = $info->{TYPE_NAME};
2656 $column_info{size} = $info->{COLUMN_SIZE};
2657 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2658 $column_info{default_value} = $info->{COLUMN_DEF};
2659 my $col_name = $info->{COLUMN_NAME};
2660 $col_name =~ s/^\"(.*)\"$/$1/;
2662 $result{$col_name} = \%column_info;
2668 return \%result if keys %result;
2671 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2674 ### The acrobatics with lc names is necessary to support both the legacy
2675 ### API that used NAME_lc exclusively, *AND* at the same time work properly
2676 ### with column names differing in cas eonly (thanks pg!)
2678 my ($columns, $seen_lcs);
2680 ++$seen_lcs->{lc($_)} and $columns->{$_} = {
2681 idx => scalar keys %$columns,
2684 } for @{$sth->{NAME}};
2686 $seen_lcs->{$_->{lc_name}} == 1
2688 $_->{name} = $_->{lc_name}
2689 for values %$columns;
2691 for ( values %$columns ) {
2693 data_type => $sth->{TYPE}->[$_->{idx}],
2694 size => $sth->{PRECISION}->[$_->{idx}],
2695 is_nullable => $sth->{NULLABLE}->[$_->{idx}] ? 1 : 0,
2698 if ($inf->{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2699 @{$inf}{qw( data_type size)} = ($1, $2);
2702 $result{$_->{name}} = $inf;
2707 if ($dbh->can('type_info')) {
2708 for my $inf (values %result) {
2709 next if ! defined $inf->{data_type};
2711 $inf->{data_type} = (
2714 $dbh->type_info( $inf->{data_type} )
2725 # FIXME - this may be an artifact of the DBD::Pg implmentation alone
2726 # needs more testing in the future...
2727 $inf->{size} -= 4 if (
2728 ( $inf->{size}||0 > 4 )
2730 $inf->{data_type} =~ qr/^text$/i
2739 sub columns_info_for {
2740 my ($self, $table) = @_;
2741 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2744 =head2 last_insert_id
2746 Return the row id of the last insert.
2750 sub _dbh_last_insert_id {
2751 my ($self, $dbh, $source, $col) = @_;
2753 my $id = dbic_internal_try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2755 return $id if defined $id;
2757 my $class = ref $self;
2758 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2761 sub last_insert_id {
2763 $self->_dbh_last_insert_id ($self->_dbh, @_);
2766 =head2 _native_data_type
2770 =item Arguments: $type_name
2774 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2775 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2776 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2778 The default implementation returns C<undef>, implement in your Storage driver if
2779 you need this functionality.
2781 Should map types from other databases to the native RDBMS type, for example
2782 C<VARCHAR2> to C<VARCHAR>.
2784 Types with modifiers should map to the underlying data type. For example,
2785 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2787 Composite types should map to the container type, for example
2788 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2792 sub _native_data_type {
2793 #my ($self, $data_type) = @_;
2797 # Check if placeholders are supported at all
2798 sub _determine_supports_placeholders {
2800 my $dbh = $self->_get_dbh;
2802 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2803 # but it is inaccurate more often than not
2804 ( dbic_internal_try {
2805 local $dbh->{PrintError} = 0;
2806 local $dbh->{RaiseError} = 1;
2807 $dbh->do('select ?', {}, 1);
2815 # Check if placeholders bound to non-string types throw exceptions
2817 sub _determine_supports_typeless_placeholders {
2819 my $dbh = $self->_get_dbh;
2821 ( dbic_internal_try {
2822 local $dbh->{PrintError} = 0;
2823 local $dbh->{RaiseError} = 1;
2824 # this specifically tests a bind that is NOT a string
2825 $dbh->do('select 1 where 1 = ?', {}, 1);
2835 Returns the database driver name.
2840 shift->_get_dbh->{Driver}->{Name};
2843 =head2 bind_attribute_by_data_type
2845 Given a datatype from column info, returns a database specific bind
2846 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2847 let the database planner just handle it.
2849 This method is always called after the driver has been determined and a DBI
2850 connection has been established. Therefore you can refer to C<DBI::$constant>
2851 and/or C<DBD::$driver::$constant> directly, without worrying about loading
2852 the correct modules.
2856 sub bind_attribute_by_data_type {
2860 =head2 is_datatype_numeric
2862 Given a datatype from column_info, returns a boolean value indicating if
2863 the current RDBMS considers it a numeric value. This controls how
2864 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2865 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2866 be performed instead of the usual C<eq>.
2870 sub is_datatype_numeric {
2871 #my ($self, $dt) = @_;
2873 return 0 unless $_[1];
2876 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2881 =head2 create_ddl_dir
2885 =item Arguments: $schema, \@databases, $version, $directory, $preversion, \%sqlt_args
2889 Creates a SQL file based on the Schema, for each of the specified
2890 database engines in C<\@databases> in the given directory.
2891 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2893 Given a previous version number, this will also create a file containing
2894 the ALTER TABLE statements to transform the previous schema into the
2895 current one. Note that these statements may contain C<DROP TABLE> or
2896 C<DROP COLUMN> statements that can potentially destroy data.
2898 The file names are created using the C<ddl_filename> method below, please
2899 override this method in your schema if you would like a different file
2900 name format. For the ALTER file, the same format is used, replacing
2901 $version in the name with "$preversion-$version".
2903 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2904 The most common value for this would be C<< { add_drop_table => 1 } >>
2905 to have the SQL produced include a C<DROP TABLE> statement for each table
2906 created. For quoting purposes supply C<quote_identifiers>.
2908 If no arguments are passed, then the following default values are assumed:
2912 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2914 =item version - $schema->schema_version
2916 =item directory - './'
2918 =item preversion - <none>
2922 By default, C<\%sqlt_args> will have
2924 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2926 merged with the hash passed in. To disable any of those features, pass in a
2927 hashref like the following
2929 { ignore_constraint_names => 0, # ... other options }
2932 WARNING: You are strongly advised to check all SQL files created, before applying
2937 sub create_ddl_dir {
2938 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2941 carp "No directory given, using ./\n";
2946 (require File::Path and File::Path::mkpath (["$dir"])) # mkpath does not like objects (i.e. Path::Class::Dir)
2948 $self->throw_exception(
2949 "Failed to create '$dir': " . ($! || $@ || 'error unknown')
2953 $self->throw_exception ("Directory '$dir' does not exist\n") unless(-d $dir);
2955 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2956 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2958 my $schema_version = $schema->schema_version || '1.x';
2959 $version ||= $schema_version;
2962 add_drop_table => 1,
2963 ignore_constraint_names => 1,
2964 ignore_index_names => 1,
2965 quote_identifiers => $self->sql_maker->_quoting_enabled,
2969 if (my $missing = DBIx::Class::Optional::Dependencies->req_missing_for ('deploy')) {
2970 $self->throw_exception("Can't create a ddl file without $missing");
2973 my $sqlt = SQL::Translator->new( $sqltargs );
2975 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2976 my $sqlt_schema = $sqlt->translate({ data => $schema })
2977 or $self->throw_exception ($sqlt->error);
2979 foreach my $db (@$databases) {
2981 $sqlt->{schema} = $sqlt_schema;
2982 $sqlt->producer($db);
2985 my $filename = $schema->ddl_filename($db, $version, $dir);
2986 if (-e $filename && ($version eq $schema_version )) {
2987 # if we are dumping the current version, overwrite the DDL
2988 carp "Overwriting existing DDL file - $filename";
2992 my $output = $sqlt->translate;
2994 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2997 if(!open($file, ">$filename")) {
2998 $self->throw_exception("Can't open $filename for writing ($!)");
3001 print $file $output;
3004 next unless ($preversion);
3006 require SQL::Translator::Diff;
3008 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
3009 if(!-e $prefilename) {
3010 carp("No previous schema file found ($prefilename)");
3014 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
3016 carp("Overwriting existing diff file - $difffile");
3022 my $t = SQL::Translator->new($sqltargs);
3027 or $self->throw_exception ($t->error);
3029 my $out = $t->translate( $prefilename )
3030 or $self->throw_exception ($t->error);
3032 $source_schema = $t->schema;
3034 $source_schema->name( $prefilename )
3035 unless ( $source_schema->name );
3038 # The "new" style of producers have sane normalization and can support
3039 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
3040 # And we have to diff parsed SQL against parsed SQL.
3041 my $dest_schema = $sqlt_schema;
3043 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
3044 my $t = SQL::Translator->new($sqltargs);
3049 or $self->throw_exception ($t->error);
3051 my $out = $t->translate( $filename )
3052 or $self->throw_exception ($t->error);
3054 $dest_schema = $t->schema;
3056 $dest_schema->name( $filename )
3057 unless $dest_schema->name;
3061 # FIXME - this is a terrible workaround for
3062 # https://github.com/dbsrgits/sql-translator/commit/2d23c1e
3063 # Fixing it in this sloppy manner so that we don't hve to
3064 # lockstep an SQLT release as well. Needs to be removed at
3065 # some point, and SQLT dep bumped
3066 local $SQL::Translator::Producer::SQLite::NO_QUOTES
3067 if $SQL::Translator::Producer::SQLite::NO_QUOTES;
3069 SQL::Translator::Diff::schema_diff($source_schema, $db,
3075 if(!open $file, ">$difffile") {
3076 $self->throw_exception("Can't write to $difffile ($!)");
3084 =head2 deployment_statements
3088 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
3092 Returns the statements used by L<DBIx::Class::Storage/deploy>
3093 and L<DBIx::Class::Schema/deploy>.
3095 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
3096 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
3098 C<$directory> is used to return statements from files in a previously created
3099 L</create_ddl_dir> directory and is optional. The filenames are constructed
3100 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
3102 If no C<$directory> is specified then the statements are constructed on the
3103 fly using L<SQL::Translator> and C<$version> is ignored.
3105 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
3109 sub deployment_statements {
3110 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
3111 $type ||= $self->sqlt_type;
3112 $version ||= $schema->schema_version || '1.x';
3114 my $filename = $schema->ddl_filename($type, $version, $dir);
3117 # FIXME replace this block when a proper sane sql parser is available
3119 open($file, "<$filename")
3120 or $self->throw_exception("Can't open $filename ($!)");
3123 return join('', @rows);
3126 if (my $missing = DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') ) {
3127 $self->throw_exception("Can't deploy without a pregenerated 'ddl_dir' directory or $missing");
3130 # sources needs to be a parser arg, but for simplicity allow at top level
3132 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
3133 if exists $sqltargs->{sources};
3135 $sqltargs->{quote_identifiers} = $self->sql_maker->_quoting_enabled
3136 unless exists $sqltargs->{quote_identifiers};
3138 my $tr = SQL::Translator->new(
3139 producer => "SQL::Translator::Producer::${type}",
3141 parser => 'SQL::Translator::Parser::DBIx::Class',
3145 return preserve_context {
3148 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
3149 unless defined $_[0];
3153 # FIXME deploy() currently does not accurately report sql errors
3154 # Will always return true while errors are warned
3156 my ($self, $schema, $type, $sqltargs, $dir) = @_;
3160 return if($line =~ /^--/);
3161 # next if($line =~ /^DROP/m);
3162 return if($line =~ /^BEGIN TRANSACTION/m);
3163 return if($line =~ /^COMMIT/m);
3164 return if $line =~ /^\s+$/; # skip whitespace only
3165 $self->_query_start($line);
3167 # do a dbh_do cycle here, as we need some error checking in
3168 # place (even though we will ignore errors)
3169 $self->dbh_do (sub { $_[1]->do($line) });
3171 carp qq{$_ (running "${line}")};
3173 $self->_query_end($line);
3175 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
3176 if (@statements > 1) {
3177 foreach my $statement (@statements) {
3178 $deploy->( $statement );
3181 elsif (@statements == 1) {
3182 # split on single line comments and end of statements
3183 foreach my $line ( split(/\s*--.*\n|;\n/, $statements[0])) {
3189 =head2 datetime_parser
3191 Returns the datetime parser class
3195 sub datetime_parser {
3197 return $self->{datetime_parser} ||= do {
3198 $self->build_datetime_parser(@_);
3202 =head2 datetime_parser_type
3204 Defines the datetime parser class - currently defaults to L<DateTime::Format::MySQL>
3206 =head2 build_datetime_parser
3208 See L</datetime_parser>
3212 sub build_datetime_parser {
3214 my $type = $self->datetime_parser_type(@_);
3219 =head2 is_replicating
3221 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
3222 replicate from a master database. Default is undef, which is the result
3223 returned by databases that don't support replication.
3227 sub is_replicating {
3232 =head2 lag_behind_master
3234 Returns a number that represents a certain amount of lag behind a master db
3235 when a given storage is replicating. The number is database dependent, but
3236 starts at zero and increases with the amount of lag. Default in undef
3240 sub lag_behind_master {
3244 =head2 relname_to_table_alias
3248 =item Arguments: $relname, $join_count
3250 =item Return Value: $alias
3254 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
3257 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
3258 way these aliases are named.
3260 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
3261 otherwise C<"$relname">.
3265 sub relname_to_table_alias {
3266 my ($self, $relname, $join_count) = @_;
3268 my $alias = ($join_count && $join_count > 1 ?
3269 join('_', $relname, $join_count) : $relname);
3274 # The size in bytes to use for DBI's ->bind_param_inout, this is the generic
3275 # version and it may be necessary to amend or override it for a specific storage
3276 # if such binds are necessary.
3277 sub _max_column_bytesize {
3278 my ($self, $attr) = @_;
3282 if ($attr->{sqlt_datatype}) {
3283 my $data_type = lc($attr->{sqlt_datatype});
3285 if ($attr->{sqlt_size}) {
3287 # String/sized-binary types
3288 if ($data_type =~ /^(?:
3289 l? (?:var)? char(?:acter)? (?:\s*varying)?
3291 (?:var)? binary (?:\s*varying)?
3296 $max_size = $attr->{sqlt_size};
3298 # Other charset/unicode types, assume scale of 4
3299 elsif ($data_type =~ /^(?:
3300 national \s* character (?:\s*varying)?
3309 $max_size = $attr->{sqlt_size} * 4;
3313 if (!$max_size and !$self->_is_lob_type($data_type)) {
3314 $max_size = 100 # for all other (numeric?) datatypes
3318 $max_size || $self->_dbic_connect_attributes->{LongReadLen} || $self->_get_dbh->{LongReadLen} || 8000;
3321 # Determine if a data_type is some type of BLOB
3323 my ($self, $data_type) = @_;
3324 $data_type && ($data_type =~ /lob|bfile|text|image|bytea|memo/i
3325 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary
3326 |varchar|character\s*varying|nvarchar
3327 |national\s*character\s*varying))?\z/xi);
3330 sub _is_binary_lob_type {
3331 my ($self, $data_type) = @_;
3332 $data_type && ($data_type =~ /blob|bfile|image|bytea/i
3333 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary))?\z/xi);
3336 sub _is_text_lob_type {
3337 my ($self, $data_type) = @_;
3338 $data_type && ($data_type =~ /^(?:clob|memo)\z/i
3339 || $data_type =~ /^long(?:\s+(?:varchar|character\s*varying|nvarchar
3340 |national\s*character\s*varying))\z/xi);
3343 # Determine if a data_type is some type of a binary type
3344 sub _is_binary_type {
3345 my ($self, $data_type) = @_;
3346 $data_type && ($self->_is_binary_lob_type($data_type)
3347 || $data_type =~ /(?:var)?(?:binary|bit|graphic)(?:\s*varying)?/i);
3354 =head2 DBIx::Class and AutoCommit
3356 DBIx::Class can do some wonderful magic with handling exceptions,
3357 disconnections, and transactions when you use C<< AutoCommit => 1 >>
3358 (the default) combined with L<txn_do|DBIx::Class::Storage/txn_do> for
3359 transaction support.
3361 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
3362 in an assumed transaction between commits, and you're telling us you'd
3363 like to manage that manually. A lot of the magic protections offered by
3364 this module will go away. We can't protect you from exceptions due to database
3365 disconnects because we don't know anything about how to restart your
3366 transactions. You're on your own for handling all sorts of exceptional
3367 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
3370 =head1 FURTHER QUESTIONS?
3372 Check the list of L<additional DBIC resources|DBIx::Class/GETTING HELP/SUPPORT>.
3374 =head1 COPYRIGHT AND LICENSE
3376 This module is free software L<copyright|DBIx::Class/COPYRIGHT AND LICENSE>
3377 by the L<DBIx::Class (DBIC) authors|DBIx::Class/AUTHORS>. You can
3378 redistribute it and/or modify it under the same terms as the
3379 L<DBIx::Class library|DBIx::Class/COPYRIGHT AND LICENSE>.