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
24 # default cursor class, overridable in connect_info attributes
25 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
27 __PACKAGE__->mk_group_accessors('inherited' => qw/
28 sql_limit_dialect sql_quote_char sql_name_sep
31 __PACKAGE__->mk_group_accessors('component_class' => qw/sql_maker_class datetime_parser_type/);
33 __PACKAGE__->sql_maker_class('DBIx::Class::SQLMaker');
34 __PACKAGE__->datetime_parser_type('DateTime::Format::MySQL'); # historic default
36 __PACKAGE__->sql_name_sep('.');
38 __PACKAGE__->mk_group_accessors('simple' => qw/
39 _connect_info _dbic_connect_attributes _driver_determined
40 _dbh _dbh_details _conn_pid _sql_maker _sql_maker_opts _dbh_autocommit
41 _perform_autoinc_retrieval _autoinc_supplied_for_op
44 # the values for these accessors are picked out (and deleted) from
45 # the attribute hashref passed to connect_info
46 my @storage_options = qw/
47 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
48 disable_sth_caching unsafe auto_savepoint
50 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
53 # capability definitions, using a 2-tiered accessor system
56 # A driver/user may define _use_X, which blindly without any checks says:
57 # "(do not) use this capability", (use_dbms_capability is an "inherited"
60 # If _use_X is undef, _supports_X is then queried. This is a "simple" style
61 # accessor, which in turn calls _determine_supports_X, and stores the return
62 # in a special slot on the storage object, which is wiped every time a $dbh
63 # reconnection takes place (it is not guaranteed that upon reconnection we
64 # will get the same rdbms version). _determine_supports_X does not need to
65 # exist on a driver, as we ->can for it before calling.
67 my @capabilities = (qw/
69 insert_returning_bound
78 __PACKAGE__->mk_group_accessors( dbms_capability => map { "_supports_$_" } @capabilities );
79 __PACKAGE__->mk_group_accessors( use_dbms_capability => map { "_use_$_" } (@capabilities ) );
81 # on by default, not strictly a capability (pending rewrite)
82 __PACKAGE__->_use_join_optimizer (1);
83 sub _determine_supports_join_optimizer { 1 };
85 # Each of these methods need _determine_driver called before itself
86 # in order to function reliably. We also need to separate accessors
87 # from plain old method calls, since an accessor called as a setter
88 # does *not* need the driver determination loop fired (and in fact
89 # can produce hard to find bugs, like e.g. losing on_connect_*
90 # semantics on fresh connections)
92 # The construct below is simply a parameterized around()
93 my $storage_accessor_idx = { map { $_ => 1 } qw(
100 for my $meth (keys %$storage_accessor_idx, qw(
101 deployment_statements
103 build_datetime_parser
115 with_deferred_fk_checks
117 get_use_dbms_capability
124 my $orig = __PACKAGE__->can ($meth)
125 or die "$meth is not a ::Storage::DBI method!";
127 my $possibly_a_setter = $storage_accessor_idx->{$meth} ? 1 : 0;
130 __PACKAGE__ ."::$meth", sprintf( <<'EOC', $possibly_a_setter, perlstring $meth ), { '$orig' => \$orig };
133 # if this is an actual *setter* - just set it, no need to connect
134 # and determine the driver
137 # only fire when invoked on an instance, a valid class-based invocation
138 # would e.g. be setting a default for an inherited accessor
141 ! $_[0]->{_driver_determined}
143 ! $_[0]->{_in_determine_driver}
145 # Only try to determine stuff if we have *something* that either is or can
146 # provide a DSN. Allows for bare $schema's generated with a plain ->connect()
147 # to still be marginally useful
148 $_[0]->_dbi_connect_info->[0]
150 $_[0]->_determine_driver;
152 # work around http://rt.perl.org/rt3/Public/Bug/Display.html?id=35878
153 goto $_[0]->can(%2$s) unless DBIx::Class::_ENV_::BROKEN_GOTO;
155 my $cref = $_[0]->can(%2$s);
165 DBIx::Class::Storage::DBI - DBI storage handler
169 my $schema = MySchema->connect('dbi:SQLite:my.db');
171 $schema->storage->debug(1);
173 my @stuff = $schema->storage->dbh_do(
175 my ($storage, $dbh, @args) = @_;
176 $dbh->do("DROP TABLE authors");
181 $schema->resultset('Book')->search({
182 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
187 This class represents the connection to an RDBMS via L<DBI>. See
188 L<DBIx::Class::Storage> for general information. This pod only
189 documents DBI-specific methods and behaviors.
196 my $new = shift->next::method(@_);
198 $new->_sql_maker_opts({});
199 $new->_dbh_details({});
200 $new->{_in_do_block} = 0;
202 # read below to see what this does
203 $new->_arm_global_destructor;
208 # This is hack to work around perl shooting stuff in random
209 # order on exit(). If we do not walk the remaining storage
210 # objects in an END block, there is a *small but real* chance
211 # of a fork()ed child to kill the parent's shared DBI handle,
212 # *before perl reaches the DESTROY in this package*
213 # Yes, it is ugly and effective.
214 # Additionally this registry is used by the CLONE method to
215 # make sure no handles are shared between threads
217 my %seek_and_destroy;
219 sub _arm_global_destructor {
221 # quick "garbage collection" pass - prevents the registry
222 # from slowly growing with a bunch of undef-valued keys
223 defined $seek_and_destroy{$_} or delete $seek_and_destroy{$_}
224 for keys %seek_and_destroy;
227 $seek_and_destroy{ refaddr($_[0]) } = $_[0]
230 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
231 # collected before leaving this scope. Depending on the code above, this
232 # may very well be just a preventive measure guarding future modifications
239 ! DBIx::Class::_ENV_::BROKEN_FORK
241 my @instances = grep { defined $_ } values %seek_and_destroy
243 local $?; # just in case the DBI destructor changes it somehow
245 # disarm the handle if not native to this process (see comment on top)
246 $_->_verify_pid for @instances;
249 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
250 # collected before leaving this scope. Depending on the code above, this
251 # may very well be just a preventive measure guarding future modifications
256 # As per DBI's recommendation, DBIC disconnects all handles as
257 # soon as possible (DBIC will reconnect only on demand from within
259 my @instances = grep { defined $_ } values %seek_and_destroy;
260 %seek_and_destroy = ();
266 # properly renumber existing refs
267 $_->_arm_global_destructor
270 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
271 # collected before leaving this scope. Depending on the code above, this
272 # may very well be just a preventive measure guarding future modifications
278 return if &detected_reinvoked_destructor;
280 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
282 # some databases spew warnings on implicit disconnect
283 return unless defined $_[0]->_dbh;
285 local $SIG{__WARN__} = sub {};
287 # not calling ->disconnect here - we are being destroyed - nothing to reset
289 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
290 # collected before leaving this scope. Depending on the code above, this
291 # may very well be just a preventive measure guarding future modifications
295 # handle pid changes correctly - do not destroy parent's connection
298 my $pid = $_[0]->_conn_pid;
300 if( defined $pid and $pid != $$ and my $dbh = $_[0]->_dbh ) {
301 $dbh->{InactiveDestroy} = 1;
306 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
307 # collected before leaving this scope. Depending on the code above, this
308 # may very well be just a preventive measure guarding future modifications
314 This method is normally called by L<DBIx::Class::Schema/connection>, which
315 encapsulates its argument list in an arrayref before passing them here.
317 The argument list may contain:
323 The same 4-element argument set one would normally pass to
324 L<DBI/connect>, optionally followed by
325 L<extra attributes|/DBIx::Class specific connection attributes>
326 recognized by DBIx::Class:
328 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
332 A single code reference which returns a connected
333 L<DBI database handle|DBI/connect> optionally followed by
334 L<extra attributes|/DBIx::Class specific connection attributes> recognized
337 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
341 A single hashref with all the attributes and the dsn/user/password
344 $connect_info_args = [{
352 $connect_info_args = [{
353 dbh_maker => sub { DBI->connect (...) },
358 This is particularly useful for L<Catalyst> based applications, allowing the
359 following config (L<Config::General> style):
364 dsn dbi:mysql:database=test
371 The C<dsn>/C<user>/C<password> combination can be substituted by the
372 C<dbh_maker> key whose value is a coderef that returns a connected
373 L<DBI database handle|DBI/connect>
377 Please note that the L<DBI> docs recommend that you always explicitly
378 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
379 recommends that it be set to I<1>, and that you perform transactions
380 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
381 to I<1> if you do not do explicitly set it to zero. This is the default
382 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
384 =head3 DBIx::Class specific connection attributes
386 In addition to the standard L<DBI|DBI/ATTRIBUTES COMMON TO ALL HANDLES>
387 L<connection|DBI/Database Handle Attributes> attributes, DBIx::Class recognizes
388 the following connection options. These options can be mixed in with your other
389 L<DBI> connection attributes, or placed in a separate hashref
390 (C<\%extra_attributes>) as shown above.
392 Every time C<connect_info> is invoked, any previous settings for
393 these options will be cleared before setting the new ones, regardless of
394 whether any options are specified in the new C<connect_info>.
401 Specifies things to do immediately after connecting or re-connecting to
402 the database. Its value may contain:
408 This contains one SQL statement to execute.
410 =item an array reference
412 This contains SQL statements to execute in order. Each element contains
413 a string or a code reference that returns a string.
415 =item a code reference
417 This contains some code to execute. Unlike code references within an
418 array reference, its return value is ignored.
422 =item on_disconnect_do
424 Takes arguments in the same form as L</on_connect_do> and executes them
425 immediately before disconnecting from the database.
427 Note, this only runs if you explicitly call L</disconnect> on the
430 =item on_connect_call
432 A more generalized form of L</on_connect_do> that calls the specified
433 C<connect_call_METHOD> methods in your storage driver.
435 on_connect_do => 'select 1'
439 on_connect_call => [ [ do_sql => 'select 1' ] ]
441 Its values may contain:
447 Will call the C<connect_call_METHOD> method.
449 =item a code reference
451 Will execute C<< $code->($storage) >>
453 =item an array reference
455 Each value can be a method name or code reference.
457 =item an array of arrays
459 For each array, the first item is taken to be the C<connect_call_> method name
460 or code reference, and the rest are parameters to it.
464 Some predefined storage methods you may use:
470 Executes a SQL string or a code reference that returns a SQL string. This is
471 what L</on_connect_do> and L</on_disconnect_do> use.
479 Will execute the scalar as SQL.
483 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
484 attributes hashref and bind values.
486 =item a code reference
488 Will execute C<< $code->($storage) >> and execute the return array refs as
495 Execute any statements necessary to initialize the database session to return
496 and accept datetime/timestamp values used with
497 L<DBIx::Class::InflateColumn::DateTime>.
499 Only necessary for some databases, see your specific storage driver for
500 implementation details.
504 =item on_disconnect_call
506 Takes arguments in the same form as L</on_connect_call> and executes them
507 immediately before disconnecting from the database.
509 Calls the C<disconnect_call_METHOD> methods as opposed to the
510 C<connect_call_METHOD> methods called by L</on_connect_call>.
512 Note, this only runs if you explicitly call L</disconnect> on the
515 =item disable_sth_caching
517 If set to a true value, this option will disable the caching of
518 statement handles via L<DBI/prepare_cached>.
522 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
523 default L</sql_limit_dialect> setting of the storage (if any). For a list
524 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
528 When true automatically sets L</quote_char> and L</name_sep> to the characters
529 appropriate for your particular RDBMS. This option is preferred over specifying
530 L</quote_char> directly.
534 Specifies what characters to use to quote table and column names.
536 C<quote_char> expects either a single character, in which case is it
537 is placed on either side of the table/column name, or an arrayref of length
538 2 in which case the table/column name is placed between the elements.
540 For example under MySQL you should use C<< quote_char => '`' >>, and for
541 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
545 This parameter is only useful in conjunction with C<quote_char>, and is used to
546 specify the character that separates elements (schemas, tables, columns) from
547 each other. If unspecified it defaults to the most commonly used C<.>.
551 This Storage driver normally installs its own C<HandleError>, sets
552 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
553 all database handles, including those supplied by a coderef. It does this
554 so that it can have consistent and useful error behavior.
556 If you set this option to a true value, Storage will not do its usual
557 modifications to the database handle's attributes, and instead relies on
558 the settings in your connect_info DBI options (or the values you set in
559 your connection coderef, in the case that you are connecting via coderef).
561 Note that your custom settings can cause Storage to malfunction,
562 especially if you set a C<HandleError> handler that suppresses exceptions
563 and/or disable C<RaiseError>.
567 If this option is true, L<DBIx::Class> will use savepoints when nesting
568 transactions, making it possible to recover from failure in the inner
569 transaction without having to abort all outer transactions.
573 Use this argument to supply a cursor class other than the default
574 L<DBIx::Class::Storage::DBI::Cursor>.
578 Some real-life examples of arguments to L</connect_info> and
579 L<DBIx::Class::Schema/connect>
581 # Simple SQLite connection
582 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
585 ->connect_info([ sub { DBI->connect(...) } ]);
587 # Connect via subref in hashref
589 dbh_maker => sub { DBI->connect(...) },
590 on_connect_do => 'alter session ...',
593 # A bit more complicated
600 { quote_char => q{"} },
604 # Equivalent to the previous example
610 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
614 # Same, but with hashref as argument
615 # See parse_connect_info for explanation
618 dsn => 'dbi:Pg:dbname=foo',
620 password => 'my_pg_password',
627 # Subref + DBIx::Class-specific connection options
630 sub { DBI->connect(...) },
634 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
635 disable_sth_caching => 1,
645 my ($self, $info) = @_;
647 return $self->_connect_info if !$info;
649 $self->_connect_info($info); # copy for _connect_info
651 $info = $self->_normalize_connect_info($info)
652 if ref $info eq 'ARRAY';
655 %{ $self->_default_dbi_connect_attributes || {} },
656 %{ $info->{attributes} || {} },
659 my @args = @{ $info->{arguments} };
661 if (keys %attrs and ref $args[0] ne 'CODE') {
663 'You provided explicit AutoCommit => 0 in your connection_info. '
664 . 'This is almost universally a bad idea (see the footnotes of '
665 . 'DBIx::Class::Storage::DBI for more info). If you still want to '
666 . 'do this you can set $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK} to disable '
668 ) if ! $attrs{AutoCommit} and ! $ENV{DBIC_UNSAFE_AUTOCOMMIT_OK};
670 push @args, \%attrs if keys %attrs;
673 # this is the authoritative "always an arrayref" thing fed to DBI->connect
674 # OR a single-element coderef-based $dbh factory
675 $self->_dbi_connect_info(\@args);
677 # extract the individual storage options
678 for my $storage_opt (keys %{ $info->{storage_options} }) {
679 my $value = $info->{storage_options}{$storage_opt};
681 $self->$storage_opt($value);
684 # Extract the individual sqlmaker options
686 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
687 # the new set of options
688 $self->_sql_maker(undef);
689 $self->_sql_maker_opts({});
691 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
692 my $value = $info->{sql_maker_options}{$sql_maker_opt};
694 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
698 # save attributes in a separate accessor so they are always
699 # introspectable, even in case of a CODE $dbhmaker
700 $self->_dbic_connect_attributes (\%attrs);
702 return $self->_connect_info;
705 sub _dbi_connect_info {
708 return $self->{_dbi_connect_info} = $_[0]
711 my $conninfo = $self->{_dbi_connect_info} || [];
713 # last ditch effort to grab a DSN
714 if ( ! defined $conninfo->[0] and $ENV{DBI_DSN} ) {
715 my @new_conninfo = @$conninfo;
716 $new_conninfo[0] = $ENV{DBI_DSN};
717 $conninfo = \@new_conninfo;
724 sub _normalize_connect_info {
725 my ($self, $info_arg) = @_;
728 my @args = @$info_arg; # take a shallow copy for further mutilation
730 # combine/pre-parse arguments depending on invocation style
733 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
734 %attrs = %{ $args[1] || {} };
737 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
738 %attrs = %{$args[0]};
740 if (my $code = delete $attrs{dbh_maker}) {
743 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
746 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
747 . "to the result of 'dbh_maker'",
749 join (', ', map { "'$_'" } (@ignored) ),
754 @args = delete @attrs{qw/dsn user password/};
757 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
759 % { $args[3] || {} },
760 % { $args[4] || {} },
762 @args = @args[0,1,2];
765 $info{arguments} = \@args;
767 my @storage_opts = grep exists $attrs{$_},
768 @storage_options, 'cursor_class';
770 @{ $info{storage_options} }{@storage_opts} =
771 delete @attrs{@storage_opts} if @storage_opts;
773 my @sql_maker_opts = grep exists $attrs{$_},
774 qw/limit_dialect quote_char name_sep quote_names/;
776 @{ $info{sql_maker_options} }{@sql_maker_opts} =
777 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
779 $info{attributes} = \%attrs if %attrs;
784 sub _default_dbi_connect_attributes () {
789 ShowErrorStatement => 1,
795 This method is deprecated in favour of setting via L</connect_info>.
799 =head2 on_disconnect_do
801 This method is deprecated in favour of setting via L</connect_info>.
805 sub _parse_connect_do {
806 my ($self, $type) = @_;
808 my $val = $self->$type;
809 return () if not defined $val;
814 push @res, [ 'do_sql', $val ];
815 } elsif (ref($val) eq 'CODE') {
817 } elsif (ref($val) eq 'ARRAY') {
818 push @res, map { [ 'do_sql', $_ ] } @$val;
820 $self->throw_exception("Invalid type for $type: ".ref($val));
828 Arguments: ($subref | $method_name), @extra_coderef_args?
830 Execute the given $subref or $method_name using the new exception-based
831 connection management.
833 The first two arguments will be the storage object that C<dbh_do> was called
834 on and a database handle to use. Any additional arguments will be passed
835 verbatim to the called subref as arguments 2 and onwards.
837 Using this (instead of $self->_dbh or $self->dbh) ensures correct
838 exception handling and reconnection (or failover in future subclasses).
840 Your subref should have no side-effects outside of the database, as
841 there is the potential for your subref to be partially double-executed
842 if the database connection was stale/dysfunctional.
846 my @stuff = $schema->storage->dbh_do(
848 my ($storage, $dbh, @cols) = @_;
849 my $cols = join(q{, }, @cols);
850 $dbh->selectrow_array("SELECT $cols FROM foo");
859 my $run_target = shift; # either a coderef or a method name
861 # short circuit when we know there is no need for a runner
863 # FIXME - assumption may be wrong
864 # the rationale for the txn_depth check is that if this block is a part
865 # of a larger transaction, everything up to that point is screwed anyway
866 return $self->$run_target($self->_get_dbh, @_)
867 if $self->{_in_do_block} or $self->transaction_depth;
869 # take a ref instead of a copy, to preserve @_ aliasing
870 # semantics within the coderef, but only if needed
871 # (pseudoforking doesn't like this trick much)
872 my $args = @_ ? \@_ : [];
874 DBIx::Class::Storage::BlockRunner->new(
877 retry_handler => sub {
878 $_[0]->failed_attempt_count == 1
880 ! $_[0]->storage->connected
883 $self->$run_target ($self->_get_dbh, @$args )
888 $_[0]->_get_dbh; # connects or reconnects on pid change, necessary to grab correct txn_depth
889 shift->next::method(@_);
894 Our C<disconnect> method also performs a rollback first if the
895 database is not in C<AutoCommit> mode.
902 # this physical disconnect below might very well throw
903 # in order to unambiguously reset the state - do the cleanup in guard
905 my $g = scope_guard {
907 defined( $self->_dbh )
908 and dbic_internal_try { $self->_dbh->disconnect };
911 $self->_dbh_details({});
912 $self->transaction_depth(undef);
913 $self->_dbh_autocommit(undef);
914 $self->savepoints([]);
916 # FIXME - this needs reenabling with the proper "no reset on same DSN" check
917 #$self->_sql_maker(undef); # this may also end up being different
922 $self->_do_connection_actions(disconnect_call_ => $_) for (
923 ( $self->on_disconnect_call || () ),
924 $self->_parse_connect_do ('on_disconnect_do')
927 # stops the "implicit rollback on disconnect" warning
928 $self->_exec_txn_rollback unless $self->_dbh_autocommit;
931 # Dummy NEXTSTATE ensuring the all temporaries on the stack are garbage
932 # collected before leaving this scope. Depending on the code above, this
933 # may very well be just a preventive measure guarding future modifications
937 =head2 with_deferred_fk_checks
941 =item Arguments: C<$coderef>
943 =item Return Value: The return value of $coderef
947 Storage specific method to run the code ref with FK checks deferred or
948 in MySQL's case disabled entirely.
952 # Storage subclasses should override this
953 sub with_deferred_fk_checks {
954 #my ($self, $sub) = @_;
962 =item Arguments: none
964 =item Return Value: 1|0
968 Verifies that the current database handle is active and ready to execute
969 an SQL statement (e.g. the connection did not get stale, server is still
970 answering, etc.) This method is used internally by L</dbh>.
975 return 0 unless $_[0]->_seems_connected;
978 local $_[0]->_dbh->{RaiseError} = 1;
983 sub _seems_connected {
984 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
988 $_[0]->_dbh->FETCH('Active')
992 # explicitly reset all state
998 ($_[0]->_dbh || return 0)->ping;
1001 sub ensure_connected {
1002 $_[0]->connected || ( $_[0]->_populate_dbh && 1 );
1007 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
1008 is guaranteed to be healthy by implicitly calling L</connected>, and if
1009 necessary performing a reconnection before returning. Keep in mind that this
1010 is very B<expensive> on some database engines. Consider using L</dbh_do>
1016 # maybe save a ping call
1018 ? ( $_[0]->ensure_connected and $_[0]->_dbh )
1019 : $_[0]->_populate_dbh
1023 # this is the internal "get dbh or connect (don't check)" method
1025 $_[0]->_verify_pid unless DBIx::Class::_ENV_::BROKEN_FORK;
1026 $_[0]->_dbh || $_[0]->_populate_dbh;
1029 # *DELIBERATELY* not a setter (for the time being)
1030 # Too intertwined with everything else for any kind of sanity
1034 $self->throw_exception('sql_maker() is not a setter method') if @_;
1036 unless ($self->_sql_maker) {
1037 my $sql_maker_class = $self->sql_maker_class;
1039 my %opts = %{$self->_sql_maker_opts||{}};
1041 $opts{limit_dialect}
1043 $self->sql_limit_dialect
1046 my $s_class = (ref $self) || $self;
1048 "Your storage class ($s_class) does not set sql_limit_dialect and you "
1049 . 'have not supplied an explicit limit_dialect in your connection_info. '
1050 . 'DBIC will attempt to use the GenericSubQ dialect, which works on most '
1051 . 'databases but can be (and often is) painfully slow. '
1052 . "Please file an RT ticket against '$s_class'"
1053 ) if $self->_dbi_connect_info->[0];
1059 my ($quote_char, $name_sep);
1061 if ($opts{quote_names}) {
1062 $quote_char = (delete $opts{quote_char}) || $self->sql_quote_char || do {
1063 my $s_class = (ref $self) || $self;
1065 "You requested 'quote_names' but your storage class ($s_class) does "
1066 . 'not explicitly define a default sql_quote_char and you have not '
1067 . 'supplied a quote_char as part of your connection_info. DBIC will '
1068 .q{default to the ANSI SQL standard quote '"', which works most of }
1069 . "the time. Please file an RT ticket against '$s_class'."
1075 $name_sep = (delete $opts{name_sep}) || $self->sql_name_sep;
1078 $self->_sql_maker($sql_maker_class->new(
1079 bindtype=>'columns',
1080 array_datatypes => 1,
1081 limit_dialect => $dialect,
1082 ($quote_char ? (quote_char => $quote_char) : ()),
1083 name_sep => ($name_sep || '.'),
1087 return $self->_sql_maker;
1090 # nothing to do by default
1096 # reset internal states
1097 # also in case ->connected failed we might get sent here
1100 $_[0]->_dbh($_[0]->_connect);
1102 $_[0]->_conn_pid($$) unless DBIx::Class::_ENV_::BROKEN_FORK; # on win32 these are in fact threads
1104 $_[0]->_determine_driver;
1106 # Always set the transaction depth on connect, since
1107 # there is no transaction in progress by definition
1108 $_[0]->transaction_depth( $_[0]->_dbh_autocommit ? 0 : 1 );
1110 $_[0]->_run_connection_actions unless $_[0]->{_in_determine_driver};
1115 sub _run_connection_actions {
1117 $_[0]->_do_connection_actions(connect_call_ => $_) for (
1118 ( $_[0]->on_connect_call || () ),
1119 $_[0]->_parse_connect_do ('on_connect_do'),
1125 sub set_use_dbms_capability {
1126 $_[0]->set_inherited ($_[1], $_[2]);
1129 sub get_use_dbms_capability {
1130 my ($self, $capname) = @_;
1132 my $use = $self->get_inherited ($capname);
1135 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1139 sub set_dbms_capability {
1140 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1143 sub get_dbms_capability {
1144 my ($self, $capname) = @_;
1146 my $cap = $self->_dbh_details->{capability}{$capname};
1148 unless (defined $cap) {
1149 if (my $meth = $self->can ("_determine$capname")) {
1150 $cap = $self->$meth ? 1 : 0;
1156 $self->set_dbms_capability ($capname, $cap);
1165 # FIXME - ideally this needs to be an ||= assignment, and the final
1166 # assignment at the end of this do{} should be gone entirely. However
1167 # this confuses CXSA: https://rt.cpan.org/Ticket/Display.html?id=103296
1168 $self->_dbh_details->{info} || do {
1170 # this guarantees that problematic conninfo won't be hidden
1171 # by the try{} below
1172 $self->ensure_connected;
1176 my $server_version = dbic_internal_try {
1177 $self->_get_server_version
1179 # driver determination *may* use this codepath
1180 # in which case we must rethrow
1181 $self->throw_exception($_) if $self->{_in_determine_driver};
1183 # $server_version on failure
1187 if (defined $server_version) {
1188 $info->{dbms_version} = $server_version;
1190 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1191 my @verparts = split (/\./, $numeric_version);
1197 # consider only up to 3 version parts, iff not more than 3 digits
1199 while (@verparts && @use_parts < 3) {
1200 my $p = shift @verparts;
1202 push @use_parts, $p;
1204 push @use_parts, 0 while @use_parts < 3;
1206 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1210 $self->_dbh_details->{info} = $info;
1214 sub _get_server_version {
1215 shift->_dbh_get_info('SQL_DBMS_VER');
1219 my ($self, $info) = @_;
1221 if ($info =~ /[^0-9]/) {
1222 require DBI::Const::GetInfoType;
1223 $info = $DBI::Const::GetInfoType::GetInfoType{$info};
1224 $self->throw_exception("Info type '$_[1]' not provided by DBI::Const::GetInfoType")
1225 unless defined $info;
1228 $self->_get_dbh->get_info($info);
1231 sub _describe_connection {
1232 require DBI::Const::GetInfoReturn;
1238 $drv = $self->_extract_driver_from_connect_info;
1239 $self->ensure_connected;
1242 $drv = "DBD::$drv" if $drv;
1245 DBIC_DSN => $self->_dbi_connect_info->[0],
1246 DBI_VER => DBI->VERSION,
1247 DBIC_VER => DBIx::Class->VERSION,
1248 DBIC_DRIVER => ref $self,
1251 DBD_VER => dbic_internal_try { $drv->VERSION },
1255 # try to grab data even if we never managed to connect
1256 # will cover us in cases of an oddly broken half-connect
1258 #keys %DBI::Const::GetInfoType::GetInfoType,
1260 SQL_CURSOR_COMMIT_BEHAVIOR
1261 SQL_CURSOR_ROLLBACK_BEHAVIOR
1262 SQL_CURSOR_SENSITIVITY
1263 SQL_DATA_SOURCE_NAME
1266 SQL_DEFAULT_TXN_ISOLATION
1271 SQL_EXPRESSIONS_IN_ORDERBY
1274 SQL_IDENTIFIER_QUOTE_CHAR
1275 SQL_MAX_CATALOG_NAME_LEN
1276 SQL_MAX_COLUMN_NAME_LEN
1277 SQL_MAX_IDENTIFIER_LEN
1278 SQL_MAX_TABLE_NAME_LEN
1279 SQL_MULTIPLE_ACTIVE_TXN
1280 SQL_MULT_RESULT_SETS
1281 SQL_NEED_LONG_DATA_LEN
1282 SQL_NON_NULLABLE_COLUMNS
1284 SQL_QUALIFIER_NAME_SEPARATOR
1285 SQL_QUOTED_IDENTIFIER_CASE
1287 SQL_TXN_ISOLATION_OPTION
1290 # some drivers barf on things they do not know about instead
1291 # of returning undef
1292 my $v = dbic_internal_try { $self->_dbh_get_info($inf) };
1293 next unless defined $v;
1295 #my $key = sprintf( '%s(%s)', $inf, $DBI::Const::GetInfoType::GetInfoType{$inf} );
1296 my $expl = DBI::Const::GetInfoReturn::Explain($inf, $v);
1297 $res->{$inf} = DBI::Const::GetInfoReturn::Format($inf, $v) . ( $expl ? " ($expl)" : '' );
1303 sub _determine_driver {
1306 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1307 my $started_connected = 0;
1308 local $self->{_in_determine_driver} = 1;
1310 if (ref($self) eq __PACKAGE__) {
1312 if ($self->_dbh) { # we are connected
1313 $driver = $self->_dbh->{Driver}{Name};
1314 $started_connected = 1;
1317 $driver = $self->_extract_driver_from_connect_info;
1321 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1322 if ($self->load_optional_class($storage_class)) {
1323 mro::set_mro($storage_class, 'c3');
1324 bless $self, $storage_class;
1328 $self->_warn_undetermined_driver(
1329 'This version of DBIC does not yet seem to supply a driver for '
1330 . "your particular RDBMS and/or connection method ('$driver')."
1335 $self->_warn_undetermined_driver(
1336 'Unable to extract a driver name from connect info - this '
1337 . 'should not have happened.'
1342 $self->_driver_determined(1);
1344 Class::C3->reinitialize() if DBIx::Class::_ENV_::OLD_MRO;
1346 if ($self->can('source_bind_attributes')) {
1347 $self->throw_exception(
1348 "Your storage subclass @{[ ref $self ]} provides (or inherits) the method "
1349 . 'source_bind_attributes() for which support has been removed as of Jan 2013. '
1350 . 'If you are not sure how to proceed please contact the development team via '
1351 . DBIx::Class::_ENV_::HELP_URL
1355 $self->_init; # run driver-specific initializations
1357 $self->_run_connection_actions
1358 if !$started_connected && defined $self->_dbh;
1362 sub _extract_driver_from_connect_info {
1367 # if connect_info is a CODEREF, we have no choice but to connect
1369 ref $self->_dbi_connect_info->[0]
1371 reftype $self->_dbi_connect_info->[0] eq 'CODE'
1373 $self->_populate_dbh;
1374 $drv = $self->_dbh->{Driver}{Name};
1377 # try to use dsn to not require being connected, the driver may still
1378 # force a connection later in _rebless to determine version
1379 # (dsn may not be supplied at all if all we do is make a mock-schema)
1380 ($drv) = ($self->_dbi_connect_info->[0] || '') =~ /^dbi:([^:]+):/i;
1381 $drv ||= $ENV{DBI_DRIVER};
1387 sub _determine_connector_driver {
1388 my ($self, $conn) = @_;
1390 my $dbtype = $self->_get_rdbms_name;
1393 $self->_warn_undetermined_driver(
1394 'Unable to retrieve RDBMS type (SQL_DBMS_NAME) of the engine behind your '
1395 . "$conn connector - this should not have happened."
1400 $dbtype =~ s/\W/_/gi;
1402 my $subclass = "DBIx::Class::Storage::DBI::${conn}::${dbtype}";
1403 return if $self->isa($subclass);
1405 if ($self->load_optional_class($subclass)) {
1406 bless $self, $subclass;
1410 $self->_warn_undetermined_driver(
1411 'This version of DBIC does not yet seem to supply a driver for '
1412 . "your particular RDBMS and/or connection method ('$conn/$dbtype')."
1417 sub _get_rdbms_name { shift->_dbh_get_info('SQL_DBMS_NAME') }
1419 sub _warn_undetermined_driver {
1420 my ($self, $msg) = @_;
1422 require Data::Dumper::Concise;
1424 carp_once ($msg . ' While we will attempt to continue anyway, the results '
1425 . 'are likely to be underwhelming. Please upgrade DBIC, and if this message '
1426 . "does not go away, file a bugreport including the following info:\n"
1427 . Data::Dumper::Concise::Dumper($self->_describe_connection)
1431 sub _do_connection_actions {
1432 my ($self, $method_prefix, $call, @args) = @_;
1435 if (not ref($call)) {
1436 my $method = $method_prefix . $call;
1437 $self->$method(@args);
1439 elsif (ref($call) eq 'CODE') {
1440 $self->$call(@args);
1442 elsif (ref($call) eq 'ARRAY') {
1443 if (ref($call->[0]) ne 'ARRAY') {
1444 $self->_do_connection_actions($method_prefix, $_) for @$call;
1447 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1451 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1455 if ( $method_prefix =~ /^connect/ ) {
1456 # this is an on_connect cycle - we can't just throw while leaving
1457 # a handle in an undefined state in our storage object
1458 # kill it with fire and rethrow
1460 $self->disconnect; # the $dbh is gone, but we still need to reset the rest
1461 $self->throw_exception( $_[0] );
1464 carp "Disconnect action failed: $_[0]";
1471 sub connect_call_do_sql {
1473 $self->_do_query(@_);
1476 sub disconnect_call_do_sql {
1478 $self->_do_query(@_);
1481 =head2 connect_call_datetime_setup
1483 A no-op stub method, provided so that one can always safely supply the
1484 L<connection option|/DBIx::Class specific connection attributes>
1486 on_connect_call => 'datetime_setup'
1488 This way one does not need to know in advance whether the underlying
1489 storage requires any sort of hand-holding when dealing with calendar
1494 sub connect_call_datetime_setup { 1 }
1497 my ($self, $action) = @_;
1499 if (ref $action eq 'CODE') {
1500 $action = $action->($self);
1501 $self->_do_query($_) foreach @$action;
1504 # Most debuggers expect ($sql, @bind), so we need to exclude
1505 # the attribute hash which is the second argument to $dbh->do
1506 # furthermore the bind values are usually to be presented
1507 # as named arrayref pairs, so wrap those here too
1508 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1509 my $sql = shift @do_args;
1510 my $attrs = shift @do_args;
1511 my @bind = map { [ undef, $_ ] } @do_args;
1514 $_[0]->_query_start($sql, \@bind);
1515 $_[1]->do($sql, $attrs, @do_args);
1516 $_[0]->_query_end($sql, \@bind);
1526 my $info = $self->_dbi_connect_info;
1528 $self->throw_exception("You did not provide any connection_info")
1529 unless defined $info->[0];
1531 my ($old_connect_via, $dbh);
1533 local $DBI::connect_via = 'connect' if $INC{'Apache/DBI.pm'} && $ENV{MOD_PERL};
1535 # this odd anonymous coderef dereference is in fact really
1536 # necessary to avoid the unwanted effect described in perl5
1539 # in addition the coderef itself can't reside inside the try{} block below
1540 # as it somehow triggers a leak under perl -d
1541 my $dbh_error_handler_installer = sub {
1542 weaken (my $weak_self = $_[0]);
1544 # the coderef is blessed so we can distinguish it from externally
1545 # supplied handles (which must be preserved)
1546 $_[1]->{HandleError} = bless sub {
1548 $weak_self->throw_exception("DBI Exception: $_[0]");
1551 # the handler may be invoked by something totally out of
1553 DBIx::Class::Exception->throw("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1555 }, '__DBIC__DBH__ERROR__HANDLER__';
1559 if(ref $info->[0] eq 'CODE') {
1560 $dbh = $info->[0]->();
1564 $dbh = DBI->connect(@$info);
1567 die $DBI::errstr unless $dbh;
1569 die sprintf ("%s fresh DBI handle with a *false* 'Active' attribute. "
1570 . 'This handle is disconnected as far as DBIC is concerned, and we can '
1572 ref $info->[0] eq 'CODE'
1573 ? "Connection coderef $info->[0] returned a"
1574 : 'DBI->connect($schema->storage->connect_info) resulted in a'
1575 ) unless $dbh->FETCH('Active');
1577 # sanity checks unless asked otherwise
1578 unless ($self->unsafe) {
1580 $self->throw_exception(
1581 'Refusing clobbering of {HandleError} installed on externally supplied '
1582 ."DBI handle $dbh. Either remove the handler or use the 'unsafe' attribute."
1583 ) if $dbh->{HandleError} and ref $dbh->{HandleError} ne '__DBIC__DBH__ERROR__HANDLER__';
1585 # Default via _default_dbi_connect_attributes is 1, hence it was an explicit
1586 # request, or an external handle. Complain and set anyway
1587 unless ($dbh->{RaiseError}) {
1588 carp( ref $info->[0] eq 'CODE'
1590 ? "The 'RaiseError' of the externally supplied DBI handle is set to false. "
1591 ."DBIx::Class will toggle it back to true, unless the 'unsafe' connect "
1592 .'attribute has been supplied'
1594 : 'RaiseError => 0 supplied in your connection_info, without an explicit '
1595 .'unsafe => 1. Toggling RaiseError back to true'
1598 $dbh->{RaiseError} = 1;
1601 $dbh_error_handler_installer->($self, $dbh);
1605 $self->throw_exception("DBI Connection failed: $_")
1608 $self->_dbh_autocommit($dbh->{AutoCommit});
1613 # this means we have not yet connected and do not know the AC status
1614 # (e.g. coderef $dbh), need a full-fledged connection check
1615 if (! defined $_[0]->_dbh_autocommit) {
1616 $_[0]->ensure_connected;
1618 # Otherwise simply connect or re-connect on pid changes
1623 shift->next::method(@_);
1626 sub _exec_txn_begin {
1629 # if the user is utilizing txn_do - good for him, otherwise we need to
1630 # ensure that the $dbh is healthy on BEGIN.
1631 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1632 # will be replaced by a failure of begin_work itself (which will be
1633 # then retried on reconnect)
1634 if ($self->{_in_do_block}) {
1635 $self->_dbh->begin_work;
1637 $self->dbh_do(sub { $_[1]->begin_work });
1644 $self->throw_exception("Unable to txn_commit() on a disconnected storage")
1645 unless $self->_seems_connected;
1647 # esoteric case for folks using external $dbh handles
1648 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1649 carp "Storage transaction_depth 0 does not match "
1650 ."false AutoCommit of $self->{_dbh}, attempting COMMIT anyway";
1651 $self->transaction_depth(1);
1654 $self->next::method(@_);
1656 # if AutoCommit is disabled txn_depth never goes to 0
1657 # as a new txn is started immediately on commit
1658 $self->transaction_depth(1) if (
1659 !$self->transaction_depth
1661 defined $self->_dbh_autocommit
1663 ! $self->_dbh_autocommit
1667 sub _exec_txn_commit {
1668 shift->_dbh->commit;
1674 # do a minimal connectivity check due to weird shit like
1675 # https://rt.cpan.org/Public/Bug/Display.html?id=62370
1676 $self->throw_exception("lost connection to storage")
1677 unless $self->_seems_connected;
1679 # esoteric case for folks using external $dbh handles
1680 if (! $self->transaction_depth and ! $self->_dbh->FETCH('AutoCommit') ) {
1681 carp "Storage transaction_depth 0 does not match "
1682 ."false AutoCommit of $self->{_dbh}, attempting ROLLBACK anyway";
1683 $self->transaction_depth(1);
1686 $self->next::method(@_);
1688 # if AutoCommit is disabled txn_depth never goes to 0
1689 # as a new txn is started immediately on commit
1690 $self->transaction_depth(1) if (
1691 !$self->transaction_depth
1693 defined $self->_dbh_autocommit
1695 ! $self->_dbh_autocommit
1699 sub _exec_txn_rollback {
1700 shift->_dbh->rollback;
1703 # generate the DBI-specific stubs, which then fallback to ::Storage proper
1704 quote_sub __PACKAGE__ . "::$_" => sprintf (<<'EOS', $_) for qw(svp_begin svp_release svp_rollback);
1705 $_[0]->throw_exception('Unable to %s() on a disconnected storage')
1706 unless $_[0]->_seems_connected;
1707 shift->next::method(@_);
1710 # This used to be the top-half of _execute. It was split out to make it
1711 # easier to override in NoBindVars without duping the rest. It takes up
1712 # all of _execute's args, and emits $sql, @bind.
1713 sub _prep_for_execute {
1714 #my ($self, $op, $ident, $args) = @_;
1715 return shift->_gen_sql_bind(@_)
1719 my ($self, $op, $ident, $args) = @_;
1721 my ($colinfos, $from);
1722 if ( blessed($ident) ) {
1723 $from = $ident->from;
1724 $colinfos = $ident->columns_info;
1728 ($sql, @$bind) = $self->sql_maker->$op( ($from || $ident), @$args );
1730 $bind = $self->_resolve_bindattrs(
1731 $ident, [ @{$args->[2]{bind}||[]}, @$bind ], $colinfos
1735 ! $ENV{DBIC_DT_SEARCH_OK}
1744 $_->[1]->isa('DateTime')
1747 carp_unique 'DateTime objects passed to search() are not supported '
1748 . 'properly (InflateColumn::DateTime formats and settings are not '
1749 . 'respected.) See ".. format a DateTime object for searching?" in '
1750 . 'DBIx::Class::Manual::FAQ. To disable this warning for good '
1751 . 'set $ENV{DBIC_DT_SEARCH_OK} to true'
1754 return( $sql, $bind );
1757 sub _resolve_bindattrs {
1758 my ($self, $ident, $bind, $colinfos) = @_;
1760 my $resolve_bindinfo = sub {
1761 #my $infohash = shift;
1763 $colinfos ||= { %{ $self->_resolve_column_info($ident) } };
1766 if (my $col = $_[0]->{dbic_colname}) {
1767 $ret = { %{$_[0]} };
1769 $ret->{sqlt_datatype} ||= $colinfos->{$col}{data_type}
1770 if $colinfos->{$col}{data_type};
1772 $ret->{sqlt_size} ||= $colinfos->{$col}{size}
1773 if $colinfos->{$col}{size};
1780 ( ref $_ ne 'ARRAY' or @$_ != 2 ) ? [ {}, $_ ]
1781 : ( ! defined $_->[0] ) ? [ {}, $_->[1] ]
1782 : (ref $_->[0] eq 'HASH') ? [(
1785 exists $_->[0]{dbd_attrs}
1787 $_->[0]{sqlt_datatype}
1789 : $resolve_bindinfo->($_->[0])
1792 : (ref $_->[0] eq 'SCALAR') ? [ { sqlt_datatype => ${$_->[0]} }, $_->[1] ]
1793 : [ $resolve_bindinfo->(
1794 { dbic_colname => $_->[0] }
1799 sub _format_for_trace {
1800 #my ($self, $bind) = @_;
1802 ### Turn @bind from something like this:
1803 ### ( [ "artist", 1 ], [ \%attrs, 3 ] )
1805 ### ( "'1'", "'3'" )
1808 defined( $_ && $_->[1] )
1815 my ( $self, $sql, $bind ) = @_;
1817 $self->debugobj->query_start( $sql, $self->_format_for_trace($bind) )
1822 my ( $self, $sql, $bind ) = @_;
1824 $self->debugobj->query_end( $sql, $self->_format_for_trace($bind) )
1828 sub _dbi_attrs_for_bind {
1829 #my ($self, $ident, $bind) = @_;
1833 exists $_->{dbd_attrs} ? $_->{dbd_attrs}
1835 : ! $_->{sqlt_datatype} ? undef
1839 # cache the result in the dbh_details hash, as it (usually) can not change
1840 # unless we connect to something else
1841 # FIXME: for the time being Oracle is an exception, pending a rewrite of
1843 my $cache = $_[0]->_dbh_details->{_datatype_map_cache} ||= {};
1845 $cache->{$_->{sqlt_datatype}} = $_[0]->bind_attribute_by_data_type($_->{sqlt_datatype})
1846 if ! exists $cache->{$_->{sqlt_datatype}};
1848 $cache->{$_->{sqlt_datatype}};
1850 } } map { $_->[0] } @{$_[2]} ];
1854 my ($self, $op, $ident, @args) = @_;
1856 my ($sql, $bind) = $self->_prep_for_execute($op, $ident, \@args);
1858 # not even a PID check - we do not care about the state of the _dbh.
1859 # All we need is to get the appropriate drivers loaded if they aren't
1860 # already so that the assumption in ad7c50fc26e holds
1861 $self->_populate_dbh unless $self->_dbh;
1863 $self->dbh_do( _dbh_execute => # retry over disconnects
1866 $self->_dbi_attrs_for_bind($ident, $bind),
1871 my ($self, $dbh, $sql, $bind, $bind_attrs) = @_;
1873 $self->_query_start( $sql, $bind );
1875 my $sth = $self->_bind_sth_params(
1876 $self->_prepare_sth($dbh, $sql),
1881 # Can this fail without throwing an exception anyways???
1882 my $rv = $sth->execute();
1883 $self->throw_exception(
1884 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1887 $self->_query_end( $sql, $bind );
1889 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1893 my ($self, $dbh, $sql) = @_;
1895 # 3 is the if_active parameter which avoids active sth re-use
1896 my $sth = $self->disable_sth_caching
1897 ? $dbh->prepare($sql)
1898 : $dbh->prepare_cached($sql, {}, 3);
1900 # XXX You would think RaiseError would make this impossible,
1901 # but apparently that's not true :(
1902 $self->throw_exception(
1905 sprintf( "\$dbh->prepare() of '%s' through %s failed *silently* without "
1906 .'an exception and/or setting $dbh->errstr',
1908 ? substr($sql, 0, 20) . '...'
1911 'DBD::' . $dbh->{Driver}{Name},
1918 sub _bind_sth_params {
1919 my ($self, $sth, $bind, $bind_attrs) = @_;
1921 for my $i (0 .. $#$bind) {
1922 if (ref $bind->[$i][1] eq 'SCALAR') { # any scalarrefs are assumed to be bind_inouts
1923 $sth->bind_param_inout(
1924 $i + 1, # bind params counts are 1-based
1926 $bind->[$i][0]{dbd_size} || $self->_max_column_bytesize($bind->[$i][0]), # size
1931 # FIXME SUBOPTIMAL - DBI needs fixing to always stringify regardless of DBD
1932 my $v = ( length ref $bind->[$i][1] and is_plain_value $bind->[$i][1] )
1939 # The temp-var is CRUCIAL - DO NOT REMOVE IT, breaks older DBD::SQLite RT#79576
1949 sub _prefetch_autovalues {
1950 my ($self, $source, $colinfo, $to_insert) = @_;
1953 for my $col (keys %$colinfo) {
1955 $colinfo->{$col}{auto_nextval}
1958 ! exists $to_insert->{$col}
1960 is_literal_value($to_insert->{$col})
1963 $values{$col} = $self->_sequence_fetch(
1965 ( $colinfo->{$col}{sequence} ||=
1966 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1976 my ($self, $source, $to_insert) = @_;
1978 my $col_infos = $source->columns_info;
1980 my $prefetched_values = $self->_prefetch_autovalues($source, $col_infos, $to_insert);
1982 # fuse the values, but keep a separate list of prefetched_values so that
1983 # they can be fused once again with the final return
1984 $to_insert = { %$to_insert, %$prefetched_values };
1986 # FIXME - we seem to assume undef values as non-supplied. This is wrong.
1987 # Investigate what does it take to s/defined/exists/
1988 my %pcols = map { $_ => 1 } $source->primary_columns;
1989 my (%retrieve_cols, $autoinc_supplied, $retrieve_autoinc_col);
1990 for my $col ($source->columns) {
1991 if ($col_infos->{$col}{is_auto_increment}) {
1992 $autoinc_supplied ||= 1 if defined $to_insert->{$col};
1993 $retrieve_autoinc_col ||= $col unless $autoinc_supplied;
1996 # nothing to retrieve when explicit values are supplied
1998 defined $to_insert->{$col} and ! is_literal_value($to_insert->{$col})
2001 # the 'scalar keys' is a trick to preserve the ->columns declaration order
2002 $retrieve_cols{$col} = scalar keys %retrieve_cols if (
2005 $col_infos->{$col}{retrieve_on_insert}
2009 local $self->{_autoinc_supplied_for_op} = $autoinc_supplied;
2010 local $self->{_perform_autoinc_retrieval} = $retrieve_autoinc_col;
2012 my ($sqla_opts, @ir_container);
2013 if (%retrieve_cols and $self->_use_insert_returning) {
2014 $sqla_opts->{returning_container} = \@ir_container
2015 if $self->_use_insert_returning_bound;
2017 $sqla_opts->{returning} = [
2018 sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols
2022 my ($rv, $sth) = $self->_execute('insert', $source, $to_insert, $sqla_opts);
2024 my %returned_cols = %$to_insert;
2025 if (my $retlist = $sqla_opts->{returning}) { # if IR is supported - we will get everything in one set
2027 unless( @ir_container ) {
2030 # FIXME - need to investigate why Caelum silenced this in 4d4dc518
2031 local $SIG{__WARN__} = sub {};
2033 @ir_container = $sth->fetchrow_array;
2037 # Evict the $sth from the cache in case we got here, since the finish()
2038 # is crucial, at least on older Firebirds, possibly on other engines too
2040 # It would be too complex to make this a proper subclass override,
2041 # and besides we already take the try{} penalty, adding a catch that
2042 # triggers infrequently is a no-brainer
2044 if( my $kids = $self->_dbh->{CachedKids} ) {
2045 $kids->{$_} == $sth and delete $kids->{$_}
2051 @returned_cols{@$retlist} = @ir_container if @ir_container;
2054 # pull in PK if needed and then everything else
2055 if (my @missing_pri = grep { $pcols{$_} } keys %retrieve_cols) {
2057 $self->throw_exception( "Missing primary key but Storage doesn't support last_insert_id" )
2058 unless $self->can('last_insert_id');
2060 my @pri_values = $self->last_insert_id($source, @missing_pri);
2062 $self->throw_exception( "Can't get last insert id" )
2063 unless (@pri_values == @missing_pri);
2065 @returned_cols{@missing_pri} = @pri_values;
2066 delete @retrieve_cols{@missing_pri};
2069 # if there is more left to pull
2070 if (%retrieve_cols) {
2071 $self->throw_exception(
2072 'Unable to retrieve additional columns without a Primary Key on ' . $source->source_name
2075 my @left_to_fetch = sort { $retrieve_cols{$a} <=> $retrieve_cols{$b} } keys %retrieve_cols;
2077 my $cur = DBIx::Class::ResultSet->new($source, {
2078 where => { map { $_ => $returned_cols{$_} } (keys %pcols) },
2079 select => \@left_to_fetch,
2082 @returned_cols{@left_to_fetch} = $cur->next;
2084 $self->throw_exception('Duplicate row returned for PK-search after fresh insert')
2085 if scalar $cur->next;
2089 return { %$prefetched_values, %returned_cols };
2094 'insert_bulk() should have never been exposed as a public method and '
2095 . 'calling it is depecated as of Aug 2014. If you believe having a genuine '
2096 . 'use for this method please contact the development team via '
2097 . DBIx::Class::_ENV_::HELP_URL
2100 return '0E0' unless @{$_[3]||[]};
2102 shift->_insert_bulk(@_);
2106 my ($self, $source, $cols, $data) = @_;
2108 $self->throw_exception('Calling _insert_bulk without a dataset to process makes no sense')
2109 unless @{$data||[]};
2111 my $colinfos = $source->columns_info($cols);
2113 local $self->{_autoinc_supplied_for_op} =
2114 (grep { $_->{is_auto_increment} } values %$colinfos)
2119 # get a slice type index based on first row of data
2120 # a "column" in this context may refer to more than one bind value
2121 # e.g. \[ '?, ?', [...], [...] ]
2123 # construct the value type index - a description of values types for every
2124 # per-column slice of $data:
2126 # nonexistent - nonbind literal
2128 # [] of bindattrs - resolved attribute(s) of bind(s) passed via literal+bind \[] combo
2130 # also construct the column hash to pass to the SQL generator. For plain
2131 # (non literal) values - convert the members of the first row into a
2132 # literal+bind combo, with extra positional info in the bind attr hashref.
2133 # This will allow us to match the order properly, and is so contrived
2134 # because a user-supplied literal/bind (or something else specific to a
2135 # resultsource and/or storage driver) can inject extra binds along the
2136 # way, so one can't rely on "shift positions" ordering at all. Also we
2137 # can't just hand SQLA a set of some known "values" (e.g. hashrefs that
2138 # can be later matched up by address), because we want to supply a real
2139 # value on which perhaps e.g. datatype checks will be performed
2140 my ($proto_data, $serialized_bind_type_by_col_idx);
2141 for my $col_idx (0..$#$cols) {
2142 my $colname = $cols->[$col_idx];
2143 if (ref $data->[0][$col_idx] eq 'SCALAR') {
2144 # no bind value at all - no type
2146 $proto_data->{$colname} = $data->[0][$col_idx];
2148 elsif (ref $data->[0][$col_idx] eq 'REF' and ref ${$data->[0][$col_idx]} eq 'ARRAY' ) {
2149 # repack, so we don't end up mangling the original \[]
2150 my ($sql, @bind) = @${$data->[0][$col_idx]};
2152 # normalization of user supplied stuff
2153 my $resolved_bind = $self->_resolve_bindattrs(
2154 $source, \@bind, $colinfos,
2157 # store value-less (attrs only) bind info - we will be comparing all
2158 # supplied binds against this for sanity
2159 $serialized_bind_type_by_col_idx->{$col_idx} = serialize [ map { $_->[0] } @$resolved_bind ];
2161 $proto_data->{$colname} = \[ $sql, map { [
2162 # inject slice order to use for $proto_bind construction
2163 { %{$resolved_bind->[$_][0]}, _bind_data_slice_idx => $col_idx, _literal_bind_subindex => $_+1 }
2165 $resolved_bind->[$_][1]
2170 $serialized_bind_type_by_col_idx->{$col_idx} = undef;
2172 $proto_data->{$colname} = \[ '?', [
2173 { dbic_colname => $colname, _bind_data_slice_idx => $col_idx }
2175 $data->[0][$col_idx]
2180 my ($sql, $proto_bind) = $self->_prep_for_execute (
2186 if (! @$proto_bind and keys %$serialized_bind_type_by_col_idx) {
2187 # if the bindlist is empty and we had some dynamic binds, this means the
2188 # storage ate them away (e.g. the NoBindVars component) and interpolated
2189 # them directly into the SQL. This obviously can't be good for multi-inserts
2190 $self->throw_exception('Unable to invoke fast-path insert without storage placeholder support');
2194 # FIXME - devise a flag "no babysitting" or somesuch to shut this off
2196 # use an error reporting closure for convenience (less to pass)
2197 my $bad_slice_report_cref = sub {
2198 my ($msg, $r_idx, $c_idx) = @_;
2199 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
2203 require Data::Dumper::Concise;
2204 local $Data::Dumper::Maxdepth = 5;
2205 Data::Dumper::Concise::Dumper ({
2206 map { $cols->[$_] =>
2214 for my $col_idx (0..$#$cols) {
2215 my $reference_val = $data->[0][$col_idx];
2217 for my $row_idx (1..$#$data) { # we are comparing against what we got from [0] above, hence start from 1
2218 my $val = $data->[$row_idx][$col_idx];
2220 if (! exists $serialized_bind_type_by_col_idx->{$col_idx}) { # literal no binds
2221 if (ref $val ne 'SCALAR') {
2222 $bad_slice_report_cref->(
2223 "Incorrect value (expecting SCALAR-ref \\'$$reference_val')",
2228 elsif ($$val ne $$reference_val) {
2229 $bad_slice_report_cref->(
2230 "Inconsistent literal SQL value (expecting \\'$$reference_val')",
2236 elsif (! defined $serialized_bind_type_by_col_idx->{$col_idx} ) { # regular non-literal value
2237 if (is_literal_value($val)) {
2238 $bad_slice_report_cref->("Literal SQL found where a plain bind value is expected", $row_idx, $col_idx);
2241 else { # binds from a \[], compare type and attrs
2242 if (ref $val ne 'REF' or ref $$val ne 'ARRAY') {
2243 $bad_slice_report_cref->(
2244 "Incorrect value (expecting ARRAYREF-ref \\['${$reference_val}->[0]', ... ])",
2249 # start drilling down and bail out early on identical refs
2251 $reference_val != $val
2253 $$reference_val != $$val
2255 if (${$val}->[0] ne ${$reference_val}->[0]) {
2256 $bad_slice_report_cref->(
2257 "Inconsistent literal/bind SQL (expecting \\['${$reference_val}->[0]', ... ])",
2262 # need to check the bind attrs - a bind will happen only once for
2263 # the entire dataset, so any changes further down will be ignored.
2265 $serialized_bind_type_by_col_idx->{$col_idx}
2270 @{$self->_resolve_bindattrs(
2271 $source, [ @{$$val}[1 .. $#$$val] ], $colinfos,
2275 $bad_slice_report_cref->(
2276 'Differing bind attributes on literal/bind values not supported',
2286 # neither _dbh_execute_for_fetch, nor _dbh_execute_inserts_with_no_binds
2287 # are atomic (even if execute_for_fetch is a single call). Thus a safety
2289 my $guard = $self->txn_scope_guard;
2291 $self->_query_start( $sql, @$proto_bind ? [[ {} => '__BULK_INSERT__' ]] : () );
2292 my $sth = $self->_prepare_sth($self->_dbh, $sql);
2295 # proto bind contains the information on which pieces of $data to pull
2296 # $cols is passed in only for prettier error-reporting
2297 $self->_dbh_execute_for_fetch( $source, $sth, $proto_bind, $cols, $data );
2300 # bind_param_array doesn't work if there are no binds
2301 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
2305 $self->_query_end( $sql, @$proto_bind ? [[ {} => '__BULK_INSERT__' ]] : () );
2309 return wantarray ? ($rv, $sth, @$proto_bind) : $rv;
2312 # execute_for_fetch is capable of returning data just fine (it means it
2313 # can be used for INSERT...RETURNING and UPDATE...RETURNING. Since this
2314 # is the void-populate fast-path we will just ignore this altogether
2315 # for the time being.
2316 sub _dbh_execute_for_fetch {
2317 my ($self, $source, $sth, $proto_bind, $cols, $data) = @_;
2319 # If we have any bind attributes to take care of, we will bind the
2320 # proto-bind data (which will never be used by execute_for_fetch)
2321 # However since column bindtypes are "sticky", this is sufficient
2322 # to get the DBD to apply the bindtype to all values later on
2323 my $bind_attrs = $self->_dbi_attrs_for_bind($source, $proto_bind);
2325 for my $i (0 .. $#$proto_bind) {
2327 $i+1, # DBI bind indexes are 1-based
2328 $proto_bind->[$i][1],
2330 ) if defined $bind_attrs->[$i];
2333 # At this point $data slots named in the _bind_data_slice_idx of
2334 # each piece of $proto_bind are either \[]s or plain values to be
2335 # passed in. Construct the dispensing coderef. *NOTE* the order
2336 # of $data will differ from this of the ?s in the SQL (due to
2337 # alphabetical ordering by colname). We actually do want to
2338 # preserve this behavior so that prepare_cached has a better
2339 # chance of matching on unrelated calls
2341 my $fetch_row_idx = -1; # saner loop this way
2342 my $fetch_tuple = sub {
2343 return undef if ++$fetch_row_idx > $#$data;
2346 my $v = ! defined $_->{_literal_bind_subindex}
2348 ? $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]
2350 # There are no attributes to resolve here - we already did everything
2351 # when we constructed proto_bind. However we still want to sanity-check
2352 # what the user supplied, so pass stuff through to the resolver *anyway*
2353 : $self->_resolve_bindattrs (
2354 undef, # a fake rsrc
2355 [ ${ $data->[ $fetch_row_idx ]->[ $_->{_bind_data_slice_idx} ]}->[ $_->{_literal_bind_subindex} ] ],
2356 {}, # a fake column_info bag
2360 # FIXME SUBOPTIMAL - DBI needs fixing to always stringify regardless of DBD
2361 # For the time being forcibly stringify whatever is stringifiable
2364 ( !length ref $v or ! ($vref = is_plain_value $v) ) ? $v
2365 : defined blessed( $$vref ) ? "$$vref"
2368 } map { $_->[0] } @$proto_bind ];
2371 my $tuple_status = [];
2374 $rv = $sth->execute_for_fetch(
2383 # Not all DBDs are create equal. Some throw on error, some return
2384 # an undef $rv, and some set $sth->err - try whatever we can
2385 $err = ($sth->errstr || 'UNKNOWN ERROR ($sth->errstr is unset)') if (
2388 ( !defined $rv or $sth->err )
2391 # Statement must finish even if there was an exception.
2396 $err = shift unless defined $err
2401 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
2403 $self->throw_exception("Unexpected populate error: $err")
2404 if ($i > $#$tuple_status);
2406 require Data::Dumper::Concise;
2407 $self->throw_exception(sprintf "execute_for_fetch() aborted with '%s' at populate slice:\n%s",
2408 ($tuple_status->[$i][1] || $err),
2409 Data::Dumper::Concise::Dumper( { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) } ),
2416 sub _dbh_execute_inserts_with_no_binds {
2417 my ($self, $sth, $count) = @_;
2421 my $dbh = $self->_get_dbh;
2422 local $dbh->{RaiseError} = 1;
2423 local $dbh->{PrintError} = 0;
2425 $sth->execute foreach 1..$count;
2431 # Make sure statement is finished even if there was an exception.
2436 $err = shift unless defined $err;
2439 $self->throw_exception($err) if defined $err;
2445 #my ($self, $source, @args) = @_;
2446 shift->_execute('update', @_);
2451 #my ($self, $source, @args) = @_;
2452 shift->_execute('delete', @_);
2457 $self->_execute($self->_select_args(@_));
2460 sub _select_args_to_query {
2463 $self->throw_exception(
2464 "Unable to generate limited query representation with 'software_limit' enabled"
2465 ) if ($_[3]->{software_limit} and ($_[3]->{offset} or $_[3]->{rows}) );
2467 # my ($op, $ident, $select, $cond, $rs_attrs, $rows, $offset)
2468 # = $self->_select_args($ident, $select, $cond, $attrs);
2469 my ($op, $ident, @args) =
2470 $self->_select_args(@_);
2472 # my ($sql, $prepared_bind) = $self->_gen_sql_bind($op, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
2473 my ($sql, $bind) = $self->_gen_sql_bind($op, $ident, \@args);
2475 # reuse the bind arrayref
2476 unshift @{$bind}, "($sql)";
2481 my ($self, $ident, $select, $where, $orig_attrs) = @_;
2483 # FIXME - that kind of caching would be nice to have
2484 # however currently we *may* pass the same $orig_attrs
2485 # with different ident/select/where
2486 # the whole interface needs to be rethought, since it
2487 # was centered around the flawed SQLA API. We can do
2488 # soooooo much better now. But that is also another
2491 # 'select', $orig_attrs->{!args_as_stored_at_the_end_of_this_method!}
2492 #) if $orig_attrs->{!args_as_stored_at_the_end_of_this_method!};
2494 my $sql_maker = $self->sql_maker;
2503 # MySQL actually recommends this approach. I cringe.
2504 $attrs->{rows} ||= $sql_maker->__max_int
2505 if $attrs->{offset};
2507 # see if we will need to tear the prefetch apart to satisfy group_by == select
2508 # this is *extremely tricky* to get right, I am still not sure I did
2510 my ($prefetch_needs_subquery, @limit_args);
2512 if ( $attrs->{_grouped_by_distinct} and $attrs->{collapse} ) {
2513 # we already know there is a valid group_by (we made it) and we know it is
2514 # intended to be based *only* on non-multi stuff
2515 # short circuit the group_by parsing below
2516 $prefetch_needs_subquery = 1;
2519 # The rationale is that even if we do *not* have collapse, we still
2520 # need to wrap the core grouped select/group_by in a subquery
2521 # so that databases that care about group_by/select equivalence
2522 # are happy (this includes MySQL in strict_mode)
2523 # If any of the other joined tables are referenced in the group_by
2524 # however - the user is on their own
2525 ( $prefetch_needs_subquery or ! $attrs->{_simple_passthrough_construction} )
2529 @{$attrs->{group_by}}
2531 my $grp_aliases = dbic_internal_try { # internal_try{} because $attrs->{from} may be unreadable
2532 $self->_resolve_aliastypes_from_select_args({ from => $attrs->{from}, group_by => $attrs->{group_by} })
2535 # no aliases other than our own in group_by
2536 # if there are - do not allow subquery even if limit is present
2537 $prefetch_needs_subquery = ! scalar grep { $_ ne $attrs->{alias} } keys %{ $grp_aliases->{grouping} || {} };
2539 elsif ( $attrs->{rows} && $attrs->{collapse} ) {
2540 # active collapse with a limit - that one is a no-brainer unless
2541 # overruled by a group_by above
2542 $prefetch_needs_subquery = 1;
2545 if ($prefetch_needs_subquery) {
2546 $attrs = $self->_adjust_select_args_for_complex_prefetch ($attrs);
2548 elsif (! $attrs->{software_limit} ) {
2550 $attrs->{rows} || (),
2551 $attrs->{offset} || (),
2555 # try to simplify the joinmap further (prune unreferenced type-single joins)
2557 ! $prefetch_needs_subquery # already pruned
2561 reftype $attrs->{from} eq 'ARRAY'
2563 @{$attrs->{from}} != 1
2565 ($attrs->{from}, $attrs->{_aliastypes}) = $self->_prune_unused_joins ($attrs);
2568 # FIXME this is a gross, inefficient, largely incorrect and fragile hack
2569 # during the result inflation stage we *need* to know what was the aliastype
2570 # map as sqla saw it when the final pieces of SQL were being assembled
2571 # Originally we simply carried around the entirety of $attrs, but this
2572 # resulted in resultsets that are being reused growing continuously, as
2573 # the hash in question grew deeper and deeper.
2574 # Instead hand-pick what to take with us here (we actually don't need much
2575 # at this point just the map itself)
2576 $orig_attrs->{_last_sqlmaker_alias_map} = $attrs->{_aliastypes};
2579 # my $alias2source = $self->_resolve_ident_sources ($ident);
2581 # This would be the point to deflate anything found in $attrs->{where}
2582 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2583 # expect a result object. And all we have is a resultsource (it is trivial
2584 # to extract deflator coderefs via $alias2source above).
2586 # I don't see a way forward other than changing the way deflators are
2587 # invoked, and that's just bad...
2590 return ( 'select', @{$attrs}{qw(from select where)}, $attrs, @limit_args );
2593 # Returns a counting SELECT for a simple count
2594 # query. Abstracted so that a storage could override
2595 # this to { count => 'firstcol' } or whatever makes
2596 # sense as a performance optimization
2598 #my ($self, $source, $rs_attrs) = @_;
2599 return { count => '*' };
2606 =item Arguments: $ident, $select, $condition, $attrs
2610 Handle a SQL select statement.
2616 my ($ident, $select, $condition, $attrs) = @_;
2617 return $self->cursor_class->new($self, \@_, $attrs);
2622 my ($rv, $sth, @bind) = $self->_select(@_);
2623 my @row = $sth->fetchrow_array;
2624 my @nextrow = $sth->fetchrow_array if @row;
2625 if(@row && @nextrow) {
2626 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2628 # Need to call finish() to work round broken DBDs
2633 =head2 sql_limit_dialect
2635 This is an accessor for the default SQL limit dialect used by a particular
2636 storage driver. Can be overridden by supplying an explicit L</limit_dialect>
2637 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2638 see L<DBIx::Class::SQLMaker::LimitDialects>.
2642 sub _dbh_columns_info_for {
2643 my ($self, $dbh, $table) = @_;
2647 if (! DBIx::Class::_ENV_::STRESSTEST_COLUMN_INFO_UNAWARE_STORAGE and $dbh->can('column_info')) {
2649 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2650 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2652 while ( my $info = $sth->fetchrow_hashref() ){
2654 $column_info{data_type} = $info->{TYPE_NAME};
2655 $column_info{size} = $info->{COLUMN_SIZE};
2656 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2657 $column_info{default_value} = $info->{COLUMN_DEF};
2658 my $col_name = $info->{COLUMN_NAME};
2659 $col_name =~ s/^\"(.*)\"$/$1/;
2661 $result{$col_name} = \%column_info;
2667 return \%result if keys %result;
2670 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2673 ### The acrobatics with lc names is necessary to support both the legacy
2674 ### API that used NAME_lc exclusively, *AND* at the same time work properly
2675 ### with column names differing in cas eonly (thanks pg!)
2677 my ($columns, $seen_lcs);
2679 ++$seen_lcs->{lc($_)} and $columns->{$_} = {
2680 idx => scalar keys %$columns,
2683 } for @{$sth->{NAME}};
2685 $seen_lcs->{$_->{lc_name}} == 1
2687 $_->{name} = $_->{lc_name}
2688 for values %$columns;
2690 for ( values %$columns ) {
2692 data_type => $sth->{TYPE}->[$_->{idx}],
2693 size => $sth->{PRECISION}->[$_->{idx}],
2694 is_nullable => $sth->{NULLABLE}->[$_->{idx}] ? 1 : 0,
2697 if ($inf->{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2698 @{$inf}{qw( data_type size)} = ($1, $2);
2701 $result{$_->{name}} = $inf;
2706 if ($dbh->can('type_info')) {
2707 for my $inf (values %result) {
2708 next if ! defined $inf->{data_type};
2710 $inf->{data_type} = (
2713 $dbh->type_info( $inf->{data_type} )
2724 # FIXME - this may be an artifact of the DBD::Pg implmentation alone
2725 # needs more testing in the future...
2726 $inf->{size} -= 4 if (
2727 ( $inf->{size}||0 > 4 )
2729 $inf->{data_type} =~ qr/^text$/i
2738 sub columns_info_for {
2739 my ($self, $table) = @_;
2740 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2743 =head2 last_insert_id
2745 Return the row id of the last insert.
2749 sub _dbh_last_insert_id {
2750 my ($self, $dbh, $source, $col) = @_;
2752 my $id = dbic_internal_try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2754 return $id if defined $id;
2756 my $class = ref $self;
2757 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2760 sub last_insert_id {
2762 $self->_dbh_last_insert_id ($self->_dbh, @_);
2765 =head2 _native_data_type
2769 =item Arguments: $type_name
2773 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2774 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2775 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2777 The default implementation returns C<undef>, implement in your Storage driver if
2778 you need this functionality.
2780 Should map types from other databases to the native RDBMS type, for example
2781 C<VARCHAR2> to C<VARCHAR>.
2783 Types with modifiers should map to the underlying data type. For example,
2784 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2786 Composite types should map to the container type, for example
2787 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2791 sub _native_data_type {
2792 #my ($self, $data_type) = @_;
2796 # Check if placeholders are supported at all
2797 sub _determine_supports_placeholders {
2799 my $dbh = $self->_get_dbh;
2801 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2802 # but it is inaccurate more often than not
2803 ( dbic_internal_try {
2804 local $dbh->{PrintError} = 0;
2805 local $dbh->{RaiseError} = 1;
2806 $dbh->do('select ?', {}, 1);
2814 # Check if placeholders bound to non-string types throw exceptions
2816 sub _determine_supports_typeless_placeholders {
2818 my $dbh = $self->_get_dbh;
2820 ( dbic_internal_try {
2821 local $dbh->{PrintError} = 0;
2822 local $dbh->{RaiseError} = 1;
2823 # this specifically tests a bind that is NOT a string
2824 $dbh->do('select 1 where 1 = ?', {}, 1);
2834 Returns the database driver name.
2839 shift->_get_dbh->{Driver}->{Name};
2842 =head2 bind_attribute_by_data_type
2844 Given a datatype from column info, returns a database specific bind
2845 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2846 let the database planner just handle it.
2848 This method is always called after the driver has been determined and a DBI
2849 connection has been established. Therefore you can refer to C<DBI::$constant>
2850 and/or C<DBD::$driver::$constant> directly, without worrying about loading
2851 the correct modules.
2855 sub bind_attribute_by_data_type {
2859 =head2 is_datatype_numeric
2861 Given a datatype from column_info, returns a boolean value indicating if
2862 the current RDBMS considers it a numeric value. This controls how
2863 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2864 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2865 be performed instead of the usual C<eq>.
2869 sub is_datatype_numeric {
2870 #my ($self, $dt) = @_;
2872 return 0 unless $_[1];
2875 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2880 =head2 create_ddl_dir
2884 =item Arguments: $schema, \@databases, $version, $directory, $preversion, \%sqlt_args
2888 Creates a SQL file based on the Schema, for each of the specified
2889 database engines in C<\@databases> in the given directory.
2890 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2892 Given a previous version number, this will also create a file containing
2893 the ALTER TABLE statements to transform the previous schema into the
2894 current one. Note that these statements may contain C<DROP TABLE> or
2895 C<DROP COLUMN> statements that can potentially destroy data.
2897 The file names are created using the C<ddl_filename> method below, please
2898 override this method in your schema if you would like a different file
2899 name format. For the ALTER file, the same format is used, replacing
2900 $version in the name with "$preversion-$version".
2902 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2903 The most common value for this would be C<< { add_drop_table => 1 } >>
2904 to have the SQL produced include a C<DROP TABLE> statement for each table
2905 created. For quoting purposes supply C<quote_identifiers>.
2907 If no arguments are passed, then the following default values are assumed:
2911 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2913 =item version - $schema->schema_version
2915 =item directory - './'
2917 =item preversion - <none>
2921 By default, C<\%sqlt_args> will have
2923 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2925 merged with the hash passed in. To disable any of those features, pass in a
2926 hashref like the following
2928 { ignore_constraint_names => 0, # ... other options }
2931 WARNING: You are strongly advised to check all SQL files created, before applying
2936 sub create_ddl_dir {
2937 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2939 require DBIx::Class::Optional::Dependencies;
2940 if (my $missing = DBIx::Class::Optional::Dependencies->req_missing_for ('deploy')) {
2941 $self->throw_exception("Can't create a ddl file without $missing");
2945 carp "No directory given, using ./\n";
2949 mkdir_p( $dir ) unless -d $dir;
2952 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2953 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2955 my $schema_version = $schema->schema_version || '1.x';
2956 $version ||= $schema_version;
2959 add_drop_table => 1,
2960 ignore_constraint_names => 1,
2961 ignore_index_names => 1,
2962 quote_identifiers => $self->sql_maker->_quoting_enabled,
2966 my $sqlt = SQL::Translator->new( $sqltargs );
2968 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2969 my $sqlt_schema = $sqlt->translate({ data => $schema })
2970 or $self->throw_exception ($sqlt->error);
2972 foreach my $db (@$databases) {
2974 $sqlt->{schema} = $sqlt_schema;
2975 $sqlt->producer($db);
2978 my $filename = $schema->ddl_filename($db, $version, $dir);
2979 if (-e $filename && ($version eq $schema_version )) {
2980 # if we are dumping the current version, overwrite the DDL
2981 carp "Overwriting existing DDL file - $filename";
2985 my $output = $sqlt->translate;
2987 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2990 if(!open($file, ">$filename")) {
2991 $self->throw_exception("Can't open $filename for writing ($!)");
2994 print $file $output;
2997 next unless ($preversion);
2999 require SQL::Translator::Diff;
3001 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
3002 if(!-e $prefilename) {
3003 carp("No previous schema file found ($prefilename)");
3007 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
3009 carp("Overwriting existing diff file - $difffile");
3015 my $t = SQL::Translator->new($sqltargs);
3020 or $self->throw_exception ($t->error);
3022 my $out = $t->translate( $prefilename )
3023 or $self->throw_exception ($t->error);
3025 $source_schema = $t->schema;
3027 $source_schema->name( $prefilename )
3028 unless ( $source_schema->name );
3031 # The "new" style of producers have sane normalization and can support
3032 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
3033 # And we have to diff parsed SQL against parsed SQL.
3034 my $dest_schema = $sqlt_schema;
3036 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
3037 my $t = SQL::Translator->new($sqltargs);
3042 or $self->throw_exception ($t->error);
3044 my $out = $t->translate( $filename )
3045 or $self->throw_exception ($t->error);
3047 $dest_schema = $t->schema;
3049 $dest_schema->name( $filename )
3050 unless $dest_schema->name;
3054 # FIXME - this is a terrible workaround for
3055 # https://github.com/dbsrgits/sql-translator/commit/2d23c1e
3056 # Fixing it in this sloppy manner so that we don't hve to
3057 # lockstep an SQLT release as well. Needs to be removed at
3058 # some point, and SQLT dep bumped
3059 local $SQL::Translator::Producer::SQLite::NO_QUOTES
3060 if $SQL::Translator::Producer::SQLite::NO_QUOTES;
3062 SQL::Translator::Diff::schema_diff($source_schema, $db,
3068 if(!open $file, ">$difffile") {
3069 $self->throw_exception("Can't write to $difffile ($!)");
3077 =head2 deployment_statements
3081 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
3085 Returns the statements used by L<DBIx::Class::Storage/deploy>
3086 and L<DBIx::Class::Schema/deploy>.
3088 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
3089 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
3091 C<$directory> is used to return statements from files in a previously created
3092 L</create_ddl_dir> directory and is optional. The filenames are constructed
3093 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
3095 If no C<$directory> is specified then the statements are constructed on the
3096 fly using L<SQL::Translator> and C<$version> is ignored.
3098 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
3102 sub deployment_statements {
3103 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
3104 $type ||= $self->sqlt_type;
3105 $version ||= $schema->schema_version || '1.x';
3107 my $filename = $schema->ddl_filename($type, $version, $dir);
3110 # FIXME replace this block when a proper sane sql parser is available
3112 open($file, "<$filename")
3113 or $self->throw_exception("Can't open $filename ($!)");
3116 return join('', @rows);
3119 require DBIx::Class::Optional::Dependencies;
3120 if (my $missing = DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') ) {
3121 $self->throw_exception("Can't deploy without a pregenerated 'ddl_dir' directory or $missing");
3124 # sources needs to be a parser arg, but for simplicity allow at top level
3126 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
3127 if exists $sqltargs->{sources};
3129 $sqltargs->{quote_identifiers} = $self->sql_maker->_quoting_enabled
3130 unless exists $sqltargs->{quote_identifiers};
3132 my $tr = SQL::Translator->new(
3133 producer => "SQL::Translator::Producer::${type}",
3135 parser => 'SQL::Translator::Parser::DBIx::Class',
3139 return preserve_context {
3142 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
3143 unless defined $_[0];
3147 # FIXME deploy() currently does not accurately report sql errors
3148 # Will always return true while errors are warned
3150 my ($self, $schema, $type, $sqltargs, $dir) = @_;
3154 return if($line =~ /^--/);
3155 # next if($line =~ /^DROP/m);
3156 return if($line =~ /^BEGIN TRANSACTION/m);
3157 return if($line =~ /^COMMIT/m);
3158 return if $line =~ /^\s+$/; # skip whitespace only
3159 $self->_query_start($line);
3161 # do a dbh_do cycle here, as we need some error checking in
3162 # place (even though we will ignore errors)
3163 $self->dbh_do (sub { $_[1]->do($line) });
3165 carp qq{$_ (running "${line}")};
3167 $self->_query_end($line);
3169 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
3170 if (@statements > 1) {
3171 foreach my $statement (@statements) {
3172 $deploy->( $statement );
3175 elsif (@statements == 1) {
3176 # split on single line comments and end of statements
3177 foreach my $line ( split(/\s*--.*\n|;\n/, $statements[0])) {
3183 =head2 datetime_parser
3185 Returns the datetime parser class
3189 sub datetime_parser {
3191 return $self->{datetime_parser} ||= do {
3192 $self->build_datetime_parser(@_);
3196 =head2 datetime_parser_type
3198 Defines the datetime parser class - currently defaults to L<DateTime::Format::MySQL>
3200 =head2 build_datetime_parser
3202 See L</datetime_parser>
3206 sub build_datetime_parser {
3208 my $type = $self->datetime_parser_type(@_);
3213 =head2 is_replicating
3215 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
3216 replicate from a master database. Default is undef, which is the result
3217 returned by databases that don't support replication.
3221 sub is_replicating {
3226 =head2 lag_behind_master
3228 Returns a number that represents a certain amount of lag behind a master db
3229 when a given storage is replicating. The number is database dependent, but
3230 starts at zero and increases with the amount of lag. Default in undef
3234 sub lag_behind_master {
3238 =head2 relname_to_table_alias
3242 =item Arguments: $relname, $join_count
3244 =item Return Value: $alias
3248 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
3251 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
3252 way these aliases are named.
3254 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
3255 otherwise C<"$relname">.
3259 sub relname_to_table_alias {
3260 my ($self, $relname, $join_count) = @_;
3262 my $alias = ($join_count && $join_count > 1 ?
3263 join('_', $relname, $join_count) : $relname);
3268 # The size in bytes to use for DBI's ->bind_param_inout, this is the generic
3269 # version and it may be necessary to amend or override it for a specific storage
3270 # if such binds are necessary.
3271 sub _max_column_bytesize {
3272 my ($self, $attr) = @_;
3276 if ($attr->{sqlt_datatype}) {
3277 my $data_type = lc($attr->{sqlt_datatype});
3279 if ($attr->{sqlt_size}) {
3281 # String/sized-binary types
3282 if ($data_type =~ /^(?:
3283 l? (?:var)? char(?:acter)? (?:\s*varying)?
3285 (?:var)? binary (?:\s*varying)?
3290 $max_size = $attr->{sqlt_size};
3292 # Other charset/unicode types, assume scale of 4
3293 elsif ($data_type =~ /^(?:
3294 national \s* character (?:\s*varying)?
3303 $max_size = $attr->{sqlt_size} * 4;
3307 if (!$max_size and !$self->_is_lob_type($data_type)) {
3308 $max_size = 100 # for all other (numeric?) datatypes
3312 $max_size || $self->_dbic_connect_attributes->{LongReadLen} || $self->_get_dbh->{LongReadLen} || 8000;
3315 # Determine if a data_type is some type of BLOB
3317 my ($self, $data_type) = @_;
3318 $data_type && ($data_type =~ /lob|bfile|text|image|bytea|memo/i
3319 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary
3320 |varchar|character\s*varying|nvarchar
3321 |national\s*character\s*varying))?\z/xi);
3324 sub _is_binary_lob_type {
3325 my ($self, $data_type) = @_;
3326 $data_type && ($data_type =~ /blob|bfile|image|bytea/i
3327 || $data_type =~ /^long(?:\s+(?:raw|bit\s*varying|varbit|binary))?\z/xi);
3330 sub _is_text_lob_type {
3331 my ($self, $data_type) = @_;
3332 $data_type && ($data_type =~ /^(?:clob|memo)\z/i
3333 || $data_type =~ /^long(?:\s+(?:varchar|character\s*varying|nvarchar
3334 |national\s*character\s*varying))\z/xi);
3337 # Determine if a data_type is some type of a binary type
3338 sub _is_binary_type {
3339 my ($self, $data_type) = @_;
3340 $data_type && ($self->_is_binary_lob_type($data_type)
3341 || $data_type =~ /(?:var)?(?:binary|bit|graphic)(?:\s*varying)?/i);
3348 =head2 DBIx::Class and AutoCommit
3350 DBIx::Class can do some wonderful magic with handling exceptions,
3351 disconnections, and transactions when you use C<< AutoCommit => 1 >>
3352 (the default) combined with L<txn_do|DBIx::Class::Storage/txn_do> for
3353 transaction support.
3355 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
3356 in an assumed transaction between commits, and you're telling us you'd
3357 like to manage that manually. A lot of the magic protections offered by
3358 this module will go away. We can't protect you from exceptions due to database
3359 disconnects because we don't know anything about how to restart your
3360 transactions. You're on your own for handling all sorts of exceptional
3361 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
3364 =head1 FURTHER QUESTIONS?
3366 Check the list of L<additional DBIC resources|DBIx::Class/GETTING HELP/SUPPORT>.
3368 =head1 COPYRIGHT AND LICENSE
3370 This module is free software L<copyright|DBIx::Class/COPYRIGHT AND LICENSE>
3371 by the L<DBIx::Class (DBIC) authors|DBIx::Class/AUTHORS>. You can
3372 redistribute it and/or modify it under the same terms as the
3373 L<DBIx::Class library|DBIx::Class/COPYRIGHT AND LICENSE>.