1 package DBIx::Class::Storage::DBI;
2 # -*- mode: cperl; cperl-indent-level: 2 -*-
7 use base qw/DBIx::Class::Storage::DBIHacks DBIx::Class::Storage/;
10 use Carp::Clan qw/^DBIx::Class|^Try::Tiny/;
12 use DBIx::Class::Storage::DBI::Cursor;
13 use Scalar::Util qw/refaddr weaken reftype blessed/;
14 use List::Util qw/first/;
15 use Data::Dumper::Concise 'Dumper';
16 use Sub::Name 'subname';
18 use File::Path 'make_path';
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/sql_maker_class sql_limit_dialect/);
27 __PACKAGE__->sql_maker_class('DBIx::Class::SQLMaker');
29 __PACKAGE__->mk_group_accessors('simple' => qw/
30 _connect_info _dbi_connect_info _dbic_connect_attributes _driver_determined
31 _dbh _dbh_details _conn_pid _sql_maker _sql_maker_opts
32 transaction_depth _dbh_autocommit savepoints
35 # the values for these accessors are picked out (and deleted) from
36 # the attribute hashref passed to connect_info
37 my @storage_options = qw/
38 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
39 disable_sth_caching unsafe auto_savepoint
41 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
44 # capability definitions, using a 2-tiered accessor system
47 # A driver/user may define _use_X, which blindly without any checks says:
48 # "(do not) use this capability", (use_dbms_capability is an "inherited"
51 # If _use_X is undef, _supports_X is then queried. This is a "simple" style
52 # accessor, which in turn calls _determine_supports_X, and stores the return
53 # in a special slot on the storage object, which is wiped every time a $dbh
54 # reconnection takes place (it is not guaranteed that upon reconnection we
55 # will get the same rdbms version). _determine_supports_X does not need to
56 # exist on a driver, as we ->can for it before calling.
58 my @capabilities = (qw/
60 insert_returning_bound
65 __PACKAGE__->mk_group_accessors( dbms_capability => map { "_supports_$_" } @capabilities );
66 __PACKAGE__->mk_group_accessors( use_dbms_capability => map { "_use_$_" } (@capabilities ) );
68 # on by default, not strictly a capability (pending rewrite)
69 __PACKAGE__->_use_join_optimizer (1);
70 sub _determine_supports_join_optimizer { 1 };
72 # Each of these methods need _determine_driver called before itself
73 # in order to function reliably. This is a purely DRY optimization
75 # get_(use)_dbms_capability need to be called on the correct Storage
76 # class, as _use_X may be hardcoded class-wide, and _supports_X calls
77 # _determine_supports_X which obv. needs a correct driver as well
78 my @rdbms_specific_methods = qw/
92 get_use_dbms_capability
99 for my $meth (@rdbms_specific_methods) {
101 my $orig = __PACKAGE__->can ($meth)
102 or die "$meth is not a ::Storage::DBI method!";
105 no warnings qw/redefine/;
106 *{__PACKAGE__ ."::$meth"} = subname $meth => sub {
107 if (not $_[0]->_driver_determined and not $_[0]->{_in_determine_driver}) {
108 $_[0]->_determine_driver;
110 # This for some reason crashes and burns on perl 5.8.1
111 # IFF the method ends up throwing an exception
112 #goto $_[0]->can ($meth);
114 my $cref = $_[0]->can ($meth);
124 DBIx::Class::Storage::DBI - DBI storage handler
128 my $schema = MySchema->connect('dbi:SQLite:my.db');
130 $schema->storage->debug(1);
132 my @stuff = $schema->storage->dbh_do(
134 my ($storage, $dbh, @args) = @_;
135 $dbh->do("DROP TABLE authors");
140 $schema->resultset('Book')->search({
141 written_on => $schema->storage->datetime_parser->format_datetime(DateTime->now)
146 This class represents the connection to an RDBMS via L<DBI>. See
147 L<DBIx::Class::Storage> for general information. This pod only
148 documents DBI-specific methods and behaviors.
155 my $new = shift->next::method(@_);
157 $new->transaction_depth(0);
158 $new->_sql_maker_opts({});
159 $new->_dbh_details({});
160 $new->{savepoints} = [];
161 $new->{_in_dbh_do} = 0;
162 $new->{_dbh_gen} = 0;
164 # read below to see what this does
165 $new->_arm_global_destructor;
170 # This is hack to work around perl shooting stuff in random
171 # order on exit(). If we do not walk the remaining storage
172 # objects in an END block, there is a *small but real* chance
173 # of a fork()ed child to kill the parent's shared DBI handle,
174 # *before perl reaches the DESTROY in this package*
175 # Yes, it is ugly and effective.
176 # Additionally this registry is used by the CLONE method to
177 # make sure no handles are shared between threads
179 my %seek_and_destroy;
181 sub _arm_global_destructor {
183 my $key = refaddr ($self);
184 $seek_and_destroy{$key} = $self;
185 weaken ($seek_and_destroy{$key});
189 local $?; # just in case the DBI destructor changes it somehow
191 # destroy just the object if not native to this process/thread
192 $_->_verify_pid for (grep
194 values %seek_and_destroy
199 # As per DBI's recommendation, DBIC disconnects all handles as
200 # soon as possible (DBIC will reconnect only on demand from within
202 for (values %seek_and_destroy) {
204 $_->{_dbh_gen}++; # so that existing cursors will drop as well
213 # some databases spew warnings on implicit disconnect
214 local $SIG{__WARN__} = sub {};
217 # this op is necessary, since the very last perl runtime statement
218 # triggers a global destruction shootout, and the $SIG localization
219 # may very well be destroyed before perl actually gets to do the
224 # handle pid changes correctly - do not destroy parent's connection
228 my $pid = $self->_conn_pid;
229 if( defined $pid and $pid != $$ and my $dbh = $self->_dbh ) {
230 $dbh->{InactiveDestroy} = 1;
240 This method is normally called by L<DBIx::Class::Schema/connection>, which
241 encapsulates its argument list in an arrayref before passing them here.
243 The argument list may contain:
249 The same 4-element argument set one would normally pass to
250 L<DBI/connect>, optionally followed by
251 L<extra attributes|/DBIx::Class specific connection attributes>
252 recognized by DBIx::Class:
254 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
258 A single code reference which returns a connected
259 L<DBI database handle|DBI/connect> optionally followed by
260 L<extra attributes|/DBIx::Class specific connection attributes> recognized
263 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
267 A single hashref with all the attributes and the dsn/user/password
270 $connect_info_args = [{
278 $connect_info_args = [{
279 dbh_maker => sub { DBI->connect (...) },
284 This is particularly useful for L<Catalyst> based applications, allowing the
285 following config (L<Config::General> style):
290 dsn dbi:mysql:database=test
297 The C<dsn>/C<user>/C<password> combination can be substituted by the
298 C<dbh_maker> key whose value is a coderef that returns a connected
299 L<DBI database handle|DBI/connect>
303 Please note that the L<DBI> docs recommend that you always explicitly
304 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
305 recommends that it be set to I<1>, and that you perform transactions
306 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
307 to I<1> if you do not do explicitly set it to zero. This is the default
308 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
310 =head3 DBIx::Class specific connection attributes
312 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
313 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
314 the following connection options. These options can be mixed in with your other
315 L<DBI> connection attributes, or placed in a separate hashref
316 (C<\%extra_attributes>) as shown above.
318 Every time C<connect_info> is invoked, any previous settings for
319 these options will be cleared before setting the new ones, regardless of
320 whether any options are specified in the new C<connect_info>.
327 Specifies things to do immediately after connecting or re-connecting to
328 the database. Its value may contain:
334 This contains one SQL statement to execute.
336 =item an array reference
338 This contains SQL statements to execute in order. Each element contains
339 a string or a code reference that returns a string.
341 =item a code reference
343 This contains some code to execute. Unlike code references within an
344 array reference, its return value is ignored.
348 =item on_disconnect_do
350 Takes arguments in the same form as L</on_connect_do> and executes them
351 immediately before disconnecting from the database.
353 Note, this only runs if you explicitly call L</disconnect> on the
356 =item on_connect_call
358 A more generalized form of L</on_connect_do> that calls the specified
359 C<connect_call_METHOD> methods in your storage driver.
361 on_connect_do => 'select 1'
365 on_connect_call => [ [ do_sql => 'select 1' ] ]
367 Its values may contain:
373 Will call the C<connect_call_METHOD> method.
375 =item a code reference
377 Will execute C<< $code->($storage) >>
379 =item an array reference
381 Each value can be a method name or code reference.
383 =item an array of arrays
385 For each array, the first item is taken to be the C<connect_call_> method name
386 or code reference, and the rest are parameters to it.
390 Some predefined storage methods you may use:
396 Executes a SQL string or a code reference that returns a SQL string. This is
397 what L</on_connect_do> and L</on_disconnect_do> use.
405 Will execute the scalar as SQL.
409 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
410 attributes hashref and bind values.
412 =item a code reference
414 Will execute C<< $code->($storage) >> and execute the return array refs as
421 Execute any statements necessary to initialize the database session to return
422 and accept datetime/timestamp values used with
423 L<DBIx::Class::InflateColumn::DateTime>.
425 Only necessary for some databases, see your specific storage driver for
426 implementation details.
430 =item on_disconnect_call
432 Takes arguments in the same form as L</on_connect_call> and executes them
433 immediately before disconnecting from the database.
435 Calls the C<disconnect_call_METHOD> methods as opposed to the
436 C<connect_call_METHOD> methods called by L</on_connect_call>.
438 Note, this only runs if you explicitly call L</disconnect> on the
441 =item disable_sth_caching
443 If set to a true value, this option will disable the caching of
444 statement handles via L<DBI/prepare_cached>.
448 Sets a specific SQL::Abstract::Limit-style limit dialect, overriding the
449 default L</sql_limit_dialect> setting of the storage (if any). For a list
450 of available limit dialects see L<DBIx::Class::SQLMaker::LimitDialects>.
454 Specifies what characters to use to quote table and column names.
456 C<quote_char> expects either a single character, in which case is it
457 is placed on either side of the table/column name, or an arrayref of length
458 2 in which case the table/column name is placed between the elements.
460 For example under MySQL you should use C<< quote_char => '`' >>, and for
461 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
465 This parameter is only useful in conjunction with C<quote_char>, and is used to
466 specify the character that separates elements (schemas, tables, columns) from
467 each other. If unspecified it defaults to the most commonly used C<.>.
471 This Storage driver normally installs its own C<HandleError>, sets
472 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
473 all database handles, including those supplied by a coderef. It does this
474 so that it can have consistent and useful error behavior.
476 If you set this option to a true value, Storage will not do its usual
477 modifications to the database handle's attributes, and instead relies on
478 the settings in your connect_info DBI options (or the values you set in
479 your connection coderef, in the case that you are connecting via coderef).
481 Note that your custom settings can cause Storage to malfunction,
482 especially if you set a C<HandleError> handler that suppresses exceptions
483 and/or disable C<RaiseError>.
487 If this option is true, L<DBIx::Class> will use savepoints when nesting
488 transactions, making it possible to recover from failure in the inner
489 transaction without having to abort all outer transactions.
493 Use this argument to supply a cursor class other than the default
494 L<DBIx::Class::Storage::DBI::Cursor>.
498 Some real-life examples of arguments to L</connect_info> and
499 L<DBIx::Class::Schema/connect>
501 # Simple SQLite connection
502 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
505 ->connect_info([ sub { DBI->connect(...) } ]);
507 # Connect via subref in hashref
509 dbh_maker => sub { DBI->connect(...) },
510 on_connect_do => 'alter session ...',
513 # A bit more complicated
520 { quote_char => q{"} },
524 # Equivalent to the previous example
530 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
534 # Same, but with hashref as argument
535 # See parse_connect_info for explanation
538 dsn => 'dbi:Pg:dbname=foo',
540 password => 'my_pg_password',
547 # Subref + DBIx::Class-specific connection options
550 sub { DBI->connect(...) },
554 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
555 disable_sth_caching => 1,
565 my ($self, $info) = @_;
567 return $self->_connect_info if !$info;
569 $self->_connect_info($info); # copy for _connect_info
571 $info = $self->_normalize_connect_info($info)
572 if ref $info eq 'ARRAY';
574 for my $storage_opt (keys %{ $info->{storage_options} }) {
575 my $value = $info->{storage_options}{$storage_opt};
577 $self->$storage_opt($value);
580 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
581 # the new set of options
582 $self->_sql_maker(undef);
583 $self->_sql_maker_opts({});
585 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
586 my $value = $info->{sql_maker_options}{$sql_maker_opt};
588 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
592 %{ $self->_default_dbi_connect_attributes || {} },
593 %{ $info->{attributes} || {} },
596 my @args = @{ $info->{arguments} };
598 $self->_dbi_connect_info([@args,
599 %attrs && !(ref $args[0] eq 'CODE') ? \%attrs : ()]);
602 # save attributes them in a separate accessor so they are always
603 # introspectable, even in case of a CODE $dbhmaker
604 $self->_dbic_connect_attributes (\%attrs);
606 return $self->_connect_info;
609 sub _normalize_connect_info {
610 my ($self, $info_arg) = @_;
613 my @args = @$info_arg; # take a shallow copy for further mutilation
615 # combine/pre-parse arguments depending on invocation style
618 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
619 %attrs = %{ $args[1] || {} };
622 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
623 %attrs = %{$args[0]};
625 if (my $code = delete $attrs{dbh_maker}) {
628 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
631 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
632 . "to the result of 'dbh_maker'",
634 join (', ', map { "'$_'" } (@ignored) ),
639 @args = delete @attrs{qw/dsn user password/};
642 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
644 % { $args[3] || {} },
645 % { $args[4] || {} },
647 @args = @args[0,1,2];
650 $info{arguments} = \@args;
652 my @storage_opts = grep exists $attrs{$_},
653 @storage_options, 'cursor_class';
655 @{ $info{storage_options} }{@storage_opts} =
656 delete @attrs{@storage_opts} if @storage_opts;
658 my @sql_maker_opts = grep exists $attrs{$_},
659 qw/limit_dialect quote_char name_sep/;
661 @{ $info{sql_maker_options} }{@sql_maker_opts} =
662 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
664 $info{attributes} = \%attrs if %attrs;
669 sub _default_dbi_connect_attributes {
679 This method is deprecated in favour of setting via L</connect_info>.
683 =head2 on_disconnect_do
685 This method is deprecated in favour of setting via L</connect_info>.
689 sub _parse_connect_do {
690 my ($self, $type) = @_;
692 my $val = $self->$type;
693 return () if not defined $val;
698 push @res, [ 'do_sql', $val ];
699 } elsif (ref($val) eq 'CODE') {
701 } elsif (ref($val) eq 'ARRAY') {
702 push @res, map { [ 'do_sql', $_ ] } @$val;
704 $self->throw_exception("Invalid type for $type: ".ref($val));
712 Arguments: ($subref | $method_name), @extra_coderef_args?
714 Execute the given $subref or $method_name using the new exception-based
715 connection management.
717 The first two arguments will be the storage object that C<dbh_do> was called
718 on and a database handle to use. Any additional arguments will be passed
719 verbatim to the called subref as arguments 2 and onwards.
721 Using this (instead of $self->_dbh or $self->dbh) ensures correct
722 exception handling and reconnection (or failover in future subclasses).
724 Your subref should have no side-effects outside of the database, as
725 there is the potential for your subref to be partially double-executed
726 if the database connection was stale/dysfunctional.
730 my @stuff = $schema->storage->dbh_do(
732 my ($storage, $dbh, @cols) = @_;
733 my $cols = join(q{, }, @cols);
734 $dbh->selectrow_array("SELECT $cols FROM foo");
745 my $dbh = $self->_get_dbh;
747 return $self->$code($dbh, @_)
748 if ( $self->{_in_dbh_do} || $self->{transaction_depth} );
750 local $self->{_in_dbh_do} = 1;
752 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
755 $self->$code ($dbh, @$args);
757 $self->throw_exception($_) if $self->connected;
759 # We were not connected - reconnect and retry, but let any
760 # exception fall right through this time
761 carp "Retrying $code after catching disconnected exception: $_"
762 if $ENV{DBIC_DBIRETRY_DEBUG};
764 $self->_populate_dbh;
765 $self->$code($self->_dbh, @$args);
769 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
770 # It also informs dbh_do to bypass itself while under the direction of txn_do,
771 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
776 ref $coderef eq 'CODE' or $self->throw_exception
777 ('$coderef must be a CODE reference');
779 local $self->{_in_dbh_do} = 1;
782 my $want = wantarray;
788 # take a ref instead of a copy, to preserve coderef @_ aliasing semantics
793 my $txn_start_depth = $self->transaction_depth;
795 @result = $coderef->(@$args);
797 elsif(defined $want) {
798 $result[0] = $coderef->(@$args);
804 my $delta_txn = $txn_start_depth - $self->transaction_depth;
805 if ($delta_txn == 0) {
808 elsif ($delta_txn != 1) {
809 # an off-by-one would mean we fired a rollback
810 carp "Unexpected reduction of transaction depth by $delta_txn after execution of $coderef";
816 if(! defined $exception) { return wantarray ? @result : $result[0] }
818 if($self->transaction_depth > 1 || $tried++ || $self->connected) {
819 my $rollback_exception;
820 try { $self->txn_rollback } catch { $rollback_exception = shift };
821 if(defined $rollback_exception) {
822 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
823 $self->throw_exception($exception) # propagate nested rollback
824 if $rollback_exception =~ /$exception_class/;
826 $self->throw_exception(
827 "Transaction aborted: ${exception}. "
828 . "Rollback failed: ${rollback_exception}"
831 $self->throw_exception($exception)
834 # We were not connected, and was first try - reconnect and retry
836 carp "Retrying $coderef after catching disconnected exception: $exception"
837 if $ENV{DBIC_TXNRETRY_DEBUG};
838 $self->_populate_dbh;
844 Our C<disconnect> method also performs a rollback first if the
845 database is not in C<AutoCommit> mode.
855 push @actions, ( $self->on_disconnect_call || () );
856 push @actions, $self->_parse_connect_do ('on_disconnect_do');
858 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
860 $self->_dbh_rollback unless $self->_dbh_autocommit;
862 %{ $self->_dbh->{CachedKids} } = ();
863 $self->_dbh->disconnect;
869 =head2 with_deferred_fk_checks
873 =item Arguments: C<$coderef>
875 =item Return Value: The return value of $coderef
879 Storage specific method to run the code ref with FK checks deferred or
880 in MySQL's case disabled entirely.
884 # Storage subclasses should override this
885 sub with_deferred_fk_checks {
886 my ($self, $sub) = @_;
894 =item Arguments: none
896 =item Return Value: 1|0
900 Verifies that the current database handle is active and ready to execute
901 an SQL statement (e.g. the connection did not get stale, server is still
902 answering, etc.) This method is used internally by L</dbh>.
908 return 0 unless $self->_seems_connected;
911 local $self->_dbh->{RaiseError} = 1;
916 sub _seems_connected {
921 my $dbh = $self->_dbh
924 return $dbh->FETCH('Active');
930 my $dbh = $self->_dbh or return 0;
935 sub ensure_connected {
938 unless ($self->connected) {
939 $self->_populate_dbh;
945 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
946 is guaranteed to be healthy by implicitly calling L</connected>, and if
947 necessary performing a reconnection before returning. Keep in mind that this
948 is very B<expensive> on some database engines. Consider using L</dbh_do>
956 if (not $self->_dbh) {
957 $self->_populate_dbh;
959 $self->ensure_connected;
964 # this is the internal "get dbh or connect (don't check)" method
968 $self->_populate_dbh unless $self->_dbh;
974 unless ($self->_sql_maker) {
975 my $sql_maker_class = $self->sql_maker_class;
976 $self->ensure_class_loaded ($sql_maker_class);
978 my %opts = %{$self->_sql_maker_opts||{}};
982 $self->sql_limit_dialect
985 my $s_class = (ref $self) || $self;
987 "Your storage class ($s_class) does not set sql_limit_dialect and you "
988 . 'have not supplied an explicit limit_dialect in your connection_info. '
989 . 'DBIC will attempt to use the GenericSubQ dialect, which works on most '
990 . 'databases but can be (and often is) painfully slow.'
997 $self->_sql_maker($sql_maker_class->new(
999 array_datatypes => 1,
1000 limit_dialect => $dialect,
1005 return $self->_sql_maker;
1008 # nothing to do by default
1015 my @info = @{$self->_dbi_connect_info || []};
1016 $self->_dbh(undef); # in case ->connected failed we might get sent here
1017 $self->_dbh_details({}); # reset everything we know
1019 $self->_dbh($self->_connect(@info));
1021 $self->_conn_pid($$) if $^O ne 'MSWin32'; # on win32 these are in fact threads
1023 $self->_determine_driver;
1025 # Always set the transaction depth on connect, since
1026 # there is no transaction in progress by definition
1027 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1029 $self->_run_connection_actions unless $self->{_in_determine_driver};
1032 sub _run_connection_actions {
1036 push @actions, ( $self->on_connect_call || () );
1037 push @actions, $self->_parse_connect_do ('on_connect_do');
1039 $self->_do_connection_actions(connect_call_ => $_) for @actions;
1044 sub set_use_dbms_capability {
1045 $_[0]->set_inherited ($_[1], $_[2]);
1048 sub get_use_dbms_capability {
1049 my ($self, $capname) = @_;
1051 my $use = $self->get_inherited ($capname);
1054 : do { $capname =~ s/^_use_/_supports_/; $self->get_dbms_capability ($capname) }
1058 sub set_dbms_capability {
1059 $_[0]->_dbh_details->{capability}{$_[1]} = $_[2];
1062 sub get_dbms_capability {
1063 my ($self, $capname) = @_;
1065 my $cap = $self->_dbh_details->{capability}{$capname};
1067 unless (defined $cap) {
1068 if (my $meth = $self->can ("_determine$capname")) {
1069 $cap = $self->$meth ? 1 : 0;
1075 $self->set_dbms_capability ($capname, $cap);
1085 unless ($info = $self->_dbh_details->{info}) {
1089 my $server_version = try { $self->_get_server_version };
1091 if (defined $server_version) {
1092 $info->{dbms_version} = $server_version;
1094 my ($numeric_version) = $server_version =~ /^([\d\.]+)/;
1095 my @verparts = split (/\./, $numeric_version);
1101 # consider only up to 3 version parts, iff not more than 3 digits
1103 while (@verparts && @use_parts < 3) {
1104 my $p = shift @verparts;
1106 push @use_parts, $p;
1108 push @use_parts, 0 while @use_parts < 3;
1110 $info->{normalized_dbms_version} = sprintf "%d.%03d%03d", @use_parts;
1114 $self->_dbh_details->{info} = $info;
1120 sub _get_server_version {
1121 shift->_get_dbh->get_info(18);
1124 sub _determine_driver {
1127 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
1128 my $started_connected = 0;
1129 local $self->{_in_determine_driver} = 1;
1131 if (ref($self) eq __PACKAGE__) {
1133 if ($self->_dbh) { # we are connected
1134 $driver = $self->_dbh->{Driver}{Name};
1135 $started_connected = 1;
1137 # if connect_info is a CODEREF, we have no choice but to connect
1138 if (ref $self->_dbi_connect_info->[0] &&
1139 reftype $self->_dbi_connect_info->[0] eq 'CODE') {
1140 $self->_populate_dbh;
1141 $driver = $self->_dbh->{Driver}{Name};
1144 # try to use dsn to not require being connected, the driver may still
1145 # force a connection in _rebless to determine version
1146 # (dsn may not be supplied at all if all we do is make a mock-schema)
1147 my $dsn = $self->_dbi_connect_info->[0] || $ENV{DBI_DSN} || '';
1148 ($driver) = $dsn =~ /dbi:([^:]+):/i;
1149 $driver ||= $ENV{DBI_DRIVER};
1154 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
1155 if ($self->load_optional_class($storage_class)) {
1156 mro::set_mro($storage_class, 'c3');
1157 bless $self, $storage_class;
1163 $self->_driver_determined(1);
1165 $self->_init; # run driver-specific initializations
1167 $self->_run_connection_actions
1168 if !$started_connected && defined $self->_dbh;
1172 sub _do_connection_actions {
1174 my $method_prefix = shift;
1177 if (not ref($call)) {
1178 my $method = $method_prefix . $call;
1180 } elsif (ref($call) eq 'CODE') {
1182 } elsif (ref($call) eq 'ARRAY') {
1183 if (ref($call->[0]) ne 'ARRAY') {
1184 $self->_do_connection_actions($method_prefix, $_) for @$call;
1186 $self->_do_connection_actions($method_prefix, @$_) for @$call;
1189 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1195 sub connect_call_do_sql {
1197 $self->_do_query(@_);
1200 sub disconnect_call_do_sql {
1202 $self->_do_query(@_);
1205 # override in db-specific backend when necessary
1206 sub connect_call_datetime_setup { 1 }
1209 my ($self, $action) = @_;
1211 if (ref $action eq 'CODE') {
1212 $action = $action->($self);
1213 $self->_do_query($_) foreach @$action;
1216 # Most debuggers expect ($sql, @bind), so we need to exclude
1217 # the attribute hash which is the second argument to $dbh->do
1218 # furthermore the bind values are usually to be presented
1219 # as named arrayref pairs, so wrap those here too
1220 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1221 my $sql = shift @do_args;
1222 my $attrs = shift @do_args;
1223 my @bind = map { [ undef, $_ ] } @do_args;
1225 $self->_query_start($sql, @bind);
1226 $self->_get_dbh->do($sql, $attrs, @do_args);
1227 $self->_query_end($sql, @bind);
1234 my ($self, @info) = @_;
1236 $self->throw_exception("You failed to provide any connection info")
1239 my ($old_connect_via, $dbh);
1241 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
1242 $old_connect_via = $DBI::connect_via;
1243 $DBI::connect_via = 'connect';
1247 if(ref $info[0] eq 'CODE') {
1248 $dbh = $info[0]->();
1251 $dbh = DBI->connect(@info);
1258 unless ($self->unsafe) {
1260 # this odd anonymous coderef dereference is in fact really
1261 # necessary to avoid the unwanted effect described in perl5
1264 my $weak_self = $_[0];
1267 $_[1]->{HandleError} = sub {
1269 $weak_self->throw_exception("DBI Exception: $_[0]");
1272 # the handler may be invoked by something totally out of
1274 croak ("DBI Exception (unhandled by DBIC, ::Schema GCed): $_[0]");
1279 $dbh->{ShowErrorStatement} = 1;
1280 $dbh->{RaiseError} = 1;
1281 $dbh->{PrintError} = 0;
1285 $self->throw_exception("DBI Connection failed: $_")
1288 $DBI::connect_via = $old_connect_via if $old_connect_via;
1291 $self->_dbh_autocommit($dbh->{AutoCommit});
1296 my ($self, $name) = @_;
1298 $name = $self->_svp_generate_name
1299 unless defined $name;
1301 $self->throw_exception ("You can't use savepoints outside a transaction")
1302 if $self->{transaction_depth} == 0;
1304 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1305 unless $self->can('_svp_begin');
1307 push @{ $self->{savepoints} }, $name;
1309 $self->debugobj->svp_begin($name) if $self->debug;
1311 return $self->_svp_begin($name);
1315 my ($self, $name) = @_;
1317 $self->throw_exception ("You can't use savepoints outside a transaction")
1318 if $self->{transaction_depth} == 0;
1320 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1321 unless $self->can('_svp_release');
1323 if (defined $name) {
1324 $self->throw_exception ("Savepoint '$name' does not exist")
1325 unless grep { $_ eq $name } @{ $self->{savepoints} };
1327 # Dig through the stack until we find the one we are releasing. This keeps
1328 # the stack up to date.
1331 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1333 $name = pop @{ $self->{savepoints} };
1336 $self->debugobj->svp_release($name) if $self->debug;
1338 return $self->_svp_release($name);
1342 my ($self, $name) = @_;
1344 $self->throw_exception ("You can't use savepoints outside a transaction")
1345 if $self->{transaction_depth} == 0;
1347 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1348 unless $self->can('_svp_rollback');
1350 if (defined $name) {
1351 # If they passed us a name, verify that it exists in the stack
1352 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1353 $self->throw_exception("Savepoint '$name' does not exist!");
1356 # Dig through the stack until we find the one we are releasing. This keeps
1357 # the stack up to date.
1358 while(my $s = pop(@{ $self->{savepoints} })) {
1359 last if($s eq $name);
1361 # Add the savepoint back to the stack, as a rollback doesn't remove the
1362 # named savepoint, only everything after it.
1363 push(@{ $self->{savepoints} }, $name);
1365 # We'll assume they want to rollback to the last savepoint
1366 $name = $self->{savepoints}->[-1];
1369 $self->debugobj->svp_rollback($name) if $self->debug;
1371 return $self->_svp_rollback($name);
1374 sub _svp_generate_name {
1376 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1382 # this means we have not yet connected and do not know the AC status
1383 # (e.g. coderef $dbh)
1384 if (! defined $self->_dbh_autocommit) {
1385 $self->ensure_connected;
1387 # otherwise re-connect on pid changes, so
1388 # that the txn_depth is adjusted properly
1389 # the lightweight _get_dbh is good enoug here
1390 # (only superficial handle check, no pings)
1395 if($self->transaction_depth == 0) {
1396 $self->debugobj->txn_begin()
1398 $self->_dbh_begin_work;
1400 elsif ($self->auto_savepoint) {
1403 $self->{transaction_depth}++;
1406 sub _dbh_begin_work {
1409 # if the user is utilizing txn_do - good for him, otherwise we need to
1410 # ensure that the $dbh is healthy on BEGIN.
1411 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1412 # will be replaced by a failure of begin_work itself (which will be
1413 # then retried on reconnect)
1414 if ($self->{_in_dbh_do}) {
1415 $self->_dbh->begin_work;
1417 $self->dbh_do(sub { $_[1]->begin_work });
1423 if ($self->{transaction_depth} == 1) {
1424 $self->debugobj->txn_commit()
1427 $self->{transaction_depth} = 0
1428 if $self->_dbh_autocommit;
1430 elsif($self->{transaction_depth} > 1) {
1431 $self->{transaction_depth}--;
1433 if $self->auto_savepoint;
1436 $self->throw_exception( 'Refusing to commit without a started transaction' );
1442 my $dbh = $self->_dbh
1443 or $self->throw_exception('cannot COMMIT on a disconnected handle');
1449 my $dbh = $self->_dbh;
1451 if ($self->{transaction_depth} == 1) {
1452 $self->debugobj->txn_rollback()
1454 $self->{transaction_depth} = 0
1455 if $self->_dbh_autocommit;
1456 $self->_dbh_rollback;
1458 elsif($self->{transaction_depth} > 1) {
1459 $self->{transaction_depth}--;
1460 if ($self->auto_savepoint) {
1461 $self->svp_rollback;
1466 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1470 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1472 if ($_ !~ /$exception_class/) {
1473 # ensure that a failed rollback resets the transaction depth
1474 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1477 $self->throw_exception($_)
1483 my $dbh = $self->_dbh
1484 or $self->throw_exception('cannot ROLLBACK on a disconnected handle');
1488 # This used to be the top-half of _execute. It was split out to make it
1489 # easier to override in NoBindVars without duping the rest. It takes up
1490 # all of _execute's args, and emits $sql, @bind.
1491 sub _prep_for_execute {
1492 my ($self, $op, $extra_bind, $ident, $args) = @_;
1494 if( blessed $ident && $ident->isa("DBIx::Class::ResultSource") ) {
1495 $ident = $ident->from();
1498 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1501 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1503 return ($sql, \@bind);
1507 sub _fix_bind_params {
1508 my ($self, @bind) = @_;
1510 ### Turn @bind from something like this:
1511 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1513 ### ( "'1'", "'1'", "'3'" )
1516 if ( defined( $_ && $_->[1] ) ) {
1517 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1524 my ( $self, $sql, @bind ) = @_;
1526 if ( $self->debug ) {
1527 @bind = $self->_fix_bind_params(@bind);
1529 $self->debugobj->query_start( $sql, @bind );
1534 my ( $self, $sql, @bind ) = @_;
1536 if ( $self->debug ) {
1537 @bind = $self->_fix_bind_params(@bind);
1538 $self->debugobj->query_end( $sql, @bind );
1543 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1545 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1547 $self->_query_start( $sql, @$bind );
1549 my $sth = $self->sth($sql,$op);
1551 my $placeholder_index = 1;
1553 foreach my $bound (@$bind) {
1554 my $attributes = {};
1555 my($column_name, @data) = @$bound;
1557 if ($bind_attributes) {
1558 $attributes = $bind_attributes->{$column_name}
1559 if defined $bind_attributes->{$column_name};
1562 foreach my $data (@data) {
1563 my $ref = ref $data;
1565 if ($ref and overload::Method($data, '""') ) {
1568 elsif ($ref eq 'SCALAR') { # any scalarrefs are assumed to be bind_inouts
1569 $sth->bind_param_inout(
1570 $placeholder_index++,
1572 $self->_max_column_bytesize($ident, $column_name),
1578 $sth->bind_param($placeholder_index++, $data, $attributes);
1582 # Can this fail without throwing an exception anyways???
1583 my $rv = $sth->execute();
1584 $self->throw_exception(
1585 $sth->errstr || $sth->err || 'Unknown error: execute() returned false, but error flags were not set...'
1588 $self->_query_end( $sql, @$bind );
1590 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1595 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1598 sub _prefetch_autovalues {
1599 my ($self, $source, $to_insert) = @_;
1601 my $colinfo = $source->columns_info;
1604 for my $col (keys %$colinfo) {
1606 $colinfo->{$col}{auto_nextval}
1609 ! exists $to_insert->{$col}
1611 ref $to_insert->{$col} eq 'SCALAR'
1614 $values{$col} = $self->_sequence_fetch(
1616 ( $colinfo->{$col}{sequence} ||=
1617 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source, $col)
1627 my ($self, $source, $to_insert) = @_;
1629 my $prefetched_values = $self->_prefetch_autovalues($source, $to_insert);
1632 $to_insert = { %$to_insert, %$prefetched_values };
1634 # list of primary keys we try to fetch from the database
1635 # both not-exsists and scalarrefs are considered
1637 for ($source->primary_columns) {
1638 $fetch_pks{$_} = scalar keys %fetch_pks # so we can preserve order for prettyness
1639 if ! exists $to_insert->{$_} or ref $to_insert->{$_} eq 'SCALAR';
1642 my ($sqla_opts, @ir_container);
1643 if ($self->_use_insert_returning) {
1645 # retain order as declared in the resultsource
1646 for (sort { $fetch_pks{$a} <=> $fetch_pks{$b} } keys %fetch_pks ) {
1647 push @{$sqla_opts->{returning}}, $_;
1648 $sqla_opts->{returning_container} = \@ir_container
1649 if $self->_use_insert_returning_bound;
1653 my $bind_attributes = $self->source_bind_attributes($source);
1655 my ($rv, $sth) = $self->_execute('insert' => [], $source, $bind_attributes, $to_insert, $sqla_opts);
1659 if (my $retlist = $sqla_opts->{returning}) {
1660 @ir_container = try {
1661 local $SIG{__WARN__} = sub {};
1662 my @r = $sth->fetchrow_array;
1665 } unless @ir_container;
1667 @returned_cols{@$retlist} = @ir_container if @ir_container;
1670 return { %$prefetched_values, %returned_cols };
1674 ## Currently it is assumed that all values passed will be "normal", i.e. not
1675 ## scalar refs, or at least, all the same type as the first set, the statement is
1676 ## only prepped once.
1678 my ($self, $source, $cols, $data) = @_;
1681 @colvalues{@$cols} = (0..$#$cols);
1683 for my $i (0..$#$cols) {
1684 my $first_val = $data->[0][$i];
1685 next unless ref $first_val eq 'SCALAR';
1687 $colvalues{ $cols->[$i] } = $first_val;
1690 # check for bad data and stringify stringifiable objects
1691 my $bad_slice = sub {
1692 my ($msg, $col_idx, $slice_idx) = @_;
1693 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
1697 local $Data::Dumper::Maxdepth = 1; # don't dump objects, if any
1699 map { $cols->[$_] => $data->[$slice_idx][$_] } (0 .. $#$cols)
1705 for my $datum_idx (0..$#$data) {
1706 my $datum = $data->[$datum_idx];
1708 for my $col_idx (0..$#$cols) {
1709 my $val = $datum->[$col_idx];
1710 my $sqla_bind = $colvalues{ $cols->[$col_idx] };
1711 my $is_literal_sql = (ref $sqla_bind) eq 'SCALAR';
1713 if ($is_literal_sql) {
1715 $bad_slice->('bind found where literal SQL expected', $col_idx, $datum_idx);
1717 elsif ((my $reftype = ref $val) ne 'SCALAR') {
1718 $bad_slice->("$reftype reference found where literal SQL expected",
1719 $col_idx, $datum_idx);
1721 elsif ($$val ne $$sqla_bind){
1722 $bad_slice->("inconsistent literal SQL value, expecting: '$$sqla_bind'",
1723 $col_idx, $datum_idx);
1726 elsif (my $reftype = ref $val) {
1728 if (overload::Method($val, '""')) {
1729 $datum->[$col_idx] = "".$val;
1732 $bad_slice->("$reftype reference found where bind expected",
1733 $col_idx, $datum_idx);
1739 my ($sql, $bind) = $self->_prep_for_execute (
1740 'insert', undef, $source, [\%colvalues]
1744 # if the bindlist is empty - make sure all "values" are in fact
1745 # literal scalarrefs. If not the case this means the storage ate
1746 # them away (e.g. the NoBindVars component) and interpolated them
1747 # directly into the SQL. This obviosly can't be good for multi-inserts
1749 $self->throw_exception('Cannot insert_bulk without support for placeholders')
1750 if first { ref $_ ne 'SCALAR' } values %colvalues;
1753 # neither _execute_array, nor _execute_inserts_with_no_binds are
1754 # atomic (even if _execute _array is a single call). Thus a safety
1756 my $guard = $self->txn_scope_guard;
1758 $self->_query_start( $sql, @$bind ? [ dummy => '__BULK_INSERT__' ] : () );
1759 my $sth = $self->sth($sql);
1762 #@bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1763 $self->_execute_array( $source, $sth, $bind, $cols, $data );
1766 # bind_param_array doesn't work if there are no binds
1767 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
1771 $self->_query_end( $sql, @$bind ? [ dummy => '__BULK_INSERT__' ] : () );
1775 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1778 sub _execute_array {
1779 my ($self, $source, $sth, $bind, $cols, $data, @extra) = @_;
1781 ## This must be an arrayref, else nothing works!
1782 my $tuple_status = [];
1784 ## Get the bind_attributes, if any exist
1785 my $bind_attributes = $self->source_bind_attributes($source);
1787 ## Bind the values and execute
1788 my $placeholder_index = 1;
1790 foreach my $bound (@$bind) {
1792 my $attributes = {};
1793 my ($column_name, $data_index) = @$bound;
1795 if( $bind_attributes ) {
1796 $attributes = $bind_attributes->{$column_name}
1797 if defined $bind_attributes->{$column_name};
1800 my @data = map { $_->[$data_index] } @$data;
1802 $sth->bind_param_array(
1805 (%$attributes ? $attributes : ()),
1807 $placeholder_index++;
1812 $rv = $self->_dbh_execute_array($sth, $tuple_status, @extra);
1818 # Statement must finish even if there was an exception.
1823 $err = shift unless defined $err
1827 if (! defined $err and $sth->err);
1831 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1833 $self->throw_exception("Unexpected populate error: $err")
1834 if ($i > $#$tuple_status);
1836 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1837 ($tuple_status->[$i][1] || $err),
1838 Dumper { map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols) },
1845 sub _dbh_execute_array {
1846 my ($self, $sth, $tuple_status, @extra) = @_;
1848 return $sth->execute_array({ArrayTupleStatus => $tuple_status});
1851 sub _dbh_execute_inserts_with_no_binds {
1852 my ($self, $sth, $count) = @_;
1856 my $dbh = $self->_get_dbh;
1857 local $dbh->{RaiseError} = 1;
1858 local $dbh->{PrintError} = 0;
1860 $sth->execute foreach 1..$count;
1866 # Make sure statement is finished even if there was an exception.
1871 $err = shift unless defined $err;
1874 $self->throw_exception($err) if defined $err;
1880 my ($self, $source, @args) = @_;
1882 my $bind_attrs = $self->source_bind_attributes($source);
1884 return $self->_execute('update' => [], $source, $bind_attrs, @args);
1889 my ($self, $source, @args) = @_;
1891 my $bind_attrs = $self->source_bind_attributes($source);
1893 return $self->_execute('delete' => [], $source, $bind_attrs, @args);
1896 # We were sent here because the $rs contains a complex search
1897 # which will require a subquery to select the correct rows
1898 # (i.e. joined or limited resultsets, or non-introspectable conditions)
1900 # Generating a single PK column subquery is trivial and supported
1901 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1902 # Look at _multipk_update_delete()
1903 sub _subq_update_delete {
1905 my ($rs, $op, $values) = @_;
1907 my $rsrc = $rs->result_source;
1909 # quick check if we got a sane rs on our hands
1910 my @pcols = $rsrc->_pri_cols;
1912 my $sel = $rs->_resolved_attrs->{select};
1913 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1916 join ("\x00", map { join '.', $rs->{attrs}{alias}, $_ } sort @pcols)
1918 join ("\x00", sort @$sel )
1920 $self->throw_exception (
1921 '_subq_update_delete can not be called on resultsets selecting columns other than the primary keys'
1928 $op eq 'update' ? $values : (),
1929 { $pcols[0] => { -in => $rs->as_query } },
1934 return $self->_multipk_update_delete (@_);
1938 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1939 # resultset update/delete involving subqueries. So by default resort
1940 # to simple (and inefficient) delete_all style per-row opearations,
1941 # while allowing specific storages to override this with a faster
1944 sub _multipk_update_delete {
1945 return shift->_per_row_update_delete (@_);
1948 # This is the default loop used to delete/update rows for multi PK
1949 # resultsets, and used by mysql exclusively (because it can't do anything
1952 # We do not use $row->$op style queries, because resultset update/delete
1953 # is not expected to cascade (this is what delete_all/update_all is for).
1955 # There should be no race conditions as the entire operation is rolled
1958 sub _per_row_update_delete {
1960 my ($rs, $op, $values) = @_;
1962 my $rsrc = $rs->result_source;
1963 my @pcols = $rsrc->_pri_cols;
1965 my $guard = $self->txn_scope_guard;
1967 # emulate the return value of $sth->execute for non-selects
1968 my $row_cnt = '0E0';
1970 my $subrs_cur = $rs->cursor;
1971 my @all_pk = $subrs_cur->all;
1972 for my $pks ( @all_pk) {
1975 for my $i (0.. $#pcols) {
1976 $cond->{$pcols[$i]} = $pks->[$i];
1981 $op eq 'update' ? $values : (),
1995 $self->_execute($self->_select_args(@_));
1998 sub _select_args_to_query {
2001 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $rs_attrs, $rows, $offset)
2002 # = $self->_select_args($ident, $select, $cond, $attrs);
2003 my ($op, $bind, $ident, $bind_attrs, @args) =
2004 $self->_select_args(@_);
2006 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $rs_attrs, $rows, $offset ]);
2007 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
2008 $prepared_bind ||= [];
2011 ? ($sql, $prepared_bind, $bind_attrs)
2012 : \[ "($sql)", @$prepared_bind ]
2017 my ($self, $ident, $select, $where, $attrs) = @_;
2019 my $sql_maker = $self->sql_maker;
2020 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
2027 $rs_alias && $alias2source->{$rs_alias}
2028 ? ( _rsroot_rsrc => $alias2source->{$rs_alias} )
2033 # calculate bind_attrs before possible $ident mangling
2034 my $bind_attrs = {};
2035 for my $alias (keys %$alias2source) {
2036 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
2037 for my $col (keys %$bindtypes) {
2039 my $fqcn = join ('.', $alias, $col);
2040 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
2042 # Unqialified column names are nice, but at the same time can be
2043 # rather ambiguous. What we do here is basically go along with
2044 # the loop, adding an unqualified column slot to $bind_attrs,
2045 # alongside the fully qualified name. As soon as we encounter
2046 # another column by that name (which would imply another table)
2047 # we unset the unqualified slot and never add any info to it
2048 # to avoid erroneous type binding. If this happens the users
2049 # only choice will be to fully qualify his column name
2051 if (exists $bind_attrs->{$col}) {
2052 $bind_attrs->{$col} = {};
2055 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
2060 # Sanity check the attributes (SQLMaker does it too, but
2061 # in case of a software_limit we'll never reach there)
2062 if (defined $attrs->{offset}) {
2063 $self->throw_exception('A supplied offset attribute must be a non-negative integer')
2064 if ( $attrs->{offset} =~ /\D/ or $attrs->{offset} < 0 );
2066 $attrs->{offset} ||= 0;
2068 if (defined $attrs->{rows}) {
2069 $self->throw_exception("The rows attribute must be a positive integer if present")
2070 if ( $attrs->{rows} =~ /\D/ or $attrs->{rows} <= 0 );
2072 elsif ($attrs->{offset}) {
2073 # MySQL actually recommends this approach. I cringe.
2074 $attrs->{rows} = $sql_maker->__max_int;
2079 # see if we need to tear the prefetch apart otherwise delegate the limiting to the
2080 # storage, unless software limit was requested
2083 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
2085 # grouped prefetch (to satisfy group_by == select)
2086 ( $attrs->{group_by}
2088 @{$attrs->{group_by}}
2090 $attrs->{_prefetch_selector_range}
2093 ($ident, $select, $where, $attrs)
2094 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
2096 elsif (! $attrs->{software_limit} ) {
2097 push @limit, $attrs->{rows}, $attrs->{offset};
2100 # try to simplify the joinmap further (prune unreferenced type-single joins)
2101 $ident = $self->_prune_unused_joins ($ident, $select, $where, $attrs);
2104 # This would be the point to deflate anything found in $where
2105 # (and leave $attrs->{bind} intact). Problem is - inflators historically
2106 # expect a row object. And all we have is a resultsource (it is trivial
2107 # to extract deflator coderefs via $alias2source above).
2109 # I don't see a way forward other than changing the way deflators are
2110 # invoked, and that's just bad...
2113 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $attrs, @limit);
2116 # Returns a counting SELECT for a simple count
2117 # query. Abstracted so that a storage could override
2118 # this to { count => 'firstcol' } or whatever makes
2119 # sense as a performance optimization
2121 #my ($self, $source, $rs_attrs) = @_;
2122 return { count => '*' };
2126 sub source_bind_attributes {
2127 my ($self, $source) = @_;
2129 my $bind_attributes;
2131 my $colinfo = $source->columns_info;
2133 for my $col (keys %$colinfo) {
2134 if (my $dt = $colinfo->{$col}{data_type} ) {
2135 $bind_attributes->{$col} = $self->bind_attribute_by_data_type($dt)
2139 return $bind_attributes;
2146 =item Arguments: $ident, $select, $condition, $attrs
2150 Handle a SQL select statement.
2156 my ($ident, $select, $condition, $attrs) = @_;
2157 return $self->cursor_class->new($self, \@_, $attrs);
2162 my ($rv, $sth, @bind) = $self->_select(@_);
2163 my @row = $sth->fetchrow_array;
2164 my @nextrow = $sth->fetchrow_array if @row;
2165 if(@row && @nextrow) {
2166 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2168 # Need to call finish() to work round broken DBDs
2173 =head2 sql_limit_dialect
2175 This is an accessor for the default SQL limit dialect used by a particular
2176 storage driver. Can be overriden by supplying an explicit L</limit_dialect>
2177 to L<DBIx::Class::Schema/connect>. For a list of available limit dialects
2178 see L<DBIx::Class::SQLMaker::LimitDialects>.
2184 =item Arguments: $sql
2188 Returns a L<DBI> sth (statement handle) for the supplied SQL.
2193 my ($self, $dbh, $sql) = @_;
2195 # 3 is the if_active parameter which avoids active sth re-use
2196 my $sth = $self->disable_sth_caching
2197 ? $dbh->prepare($sql)
2198 : $dbh->prepare_cached($sql, {}, 3);
2200 # XXX You would think RaiseError would make this impossible,
2201 # but apparently that's not true :(
2202 $self->throw_exception($dbh->errstr) if !$sth;
2208 my ($self, $sql) = @_;
2209 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2212 sub _dbh_columns_info_for {
2213 my ($self, $dbh, $table) = @_;
2215 if ($dbh->can('column_info')) {
2219 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2220 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2222 while ( my $info = $sth->fetchrow_hashref() ){
2224 $column_info{data_type} = $info->{TYPE_NAME};
2225 $column_info{size} = $info->{COLUMN_SIZE};
2226 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2227 $column_info{default_value} = $info->{COLUMN_DEF};
2228 my $col_name = $info->{COLUMN_NAME};
2229 $col_name =~ s/^\"(.*)\"$/$1/;
2231 $result{$col_name} = \%column_info;
2236 return \%result if !$caught && scalar keys %result;
2240 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2242 my @columns = @{$sth->{NAME_lc}};
2243 for my $i ( 0 .. $#columns ){
2245 $column_info{data_type} = $sth->{TYPE}->[$i];
2246 $column_info{size} = $sth->{PRECISION}->[$i];
2247 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2249 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2250 $column_info{data_type} = $1;
2251 $column_info{size} = $2;
2254 $result{$columns[$i]} = \%column_info;
2258 foreach my $col (keys %result) {
2259 my $colinfo = $result{$col};
2260 my $type_num = $colinfo->{data_type};
2262 if(defined $type_num && $dbh->can('type_info')) {
2263 my $type_info = $dbh->type_info($type_num);
2264 $type_name = $type_info->{TYPE_NAME} if $type_info;
2265 $colinfo->{data_type} = $type_name if $type_name;
2272 sub columns_info_for {
2273 my ($self, $table) = @_;
2274 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2277 =head2 last_insert_id
2279 Return the row id of the last insert.
2283 sub _dbh_last_insert_id {
2284 my ($self, $dbh, $source, $col) = @_;
2286 my $id = try { $dbh->last_insert_id (undef, undef, $source->name, $col) };
2288 return $id if defined $id;
2290 my $class = ref $self;
2291 $self->throw_exception ("No storage specific _dbh_last_insert_id() method implemented in $class, and the generic DBI::last_insert_id() failed");
2294 sub last_insert_id {
2296 $self->_dbh_last_insert_id ($self->_dbh, @_);
2299 =head2 _native_data_type
2303 =item Arguments: $type_name
2307 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2308 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2309 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2311 The default implementation returns C<undef>, implement in your Storage driver if
2312 you need this functionality.
2314 Should map types from other databases to the native RDBMS type, for example
2315 C<VARCHAR2> to C<VARCHAR>.
2317 Types with modifiers should map to the underlying data type. For example,
2318 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2320 Composite types should map to the container type, for example
2321 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2325 sub _native_data_type {
2326 #my ($self, $data_type) = @_;
2330 # Check if placeholders are supported at all
2331 sub _determine_supports_placeholders {
2333 my $dbh = $self->_get_dbh;
2335 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2336 # but it is inaccurate more often than not
2338 local $dbh->{PrintError} = 0;
2339 local $dbh->{RaiseError} = 1;
2340 $dbh->do('select ?', {}, 1);
2348 # Check if placeholders bound to non-string types throw exceptions
2350 sub _determine_supports_typeless_placeholders {
2352 my $dbh = $self->_get_dbh;
2355 local $dbh->{PrintError} = 0;
2356 local $dbh->{RaiseError} = 1;
2357 # this specifically tests a bind that is NOT a string
2358 $dbh->do('select 1 where 1 = ?', {}, 1);
2368 Returns the database driver name.
2373 shift->_get_dbh->{Driver}->{Name};
2376 =head2 bind_attribute_by_data_type
2378 Given a datatype from column info, returns a database specific bind
2379 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2380 let the database planner just handle it.
2382 Generally only needed for special case column types, like bytea in postgres.
2386 sub bind_attribute_by_data_type {
2390 =head2 is_datatype_numeric
2392 Given a datatype from column_info, returns a boolean value indicating if
2393 the current RDBMS considers it a numeric value. This controls how
2394 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2395 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2396 be performed instead of the usual C<eq>.
2400 sub is_datatype_numeric {
2401 my ($self, $dt) = @_;
2403 return 0 unless $dt;
2405 return $dt =~ /^ (?:
2406 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2411 =head2 create_ddl_dir
2415 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2419 Creates a SQL file based on the Schema, for each of the specified
2420 database engines in C<\@databases> in the given directory.
2421 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2423 Given a previous version number, this will also create a file containing
2424 the ALTER TABLE statements to transform the previous schema into the
2425 current one. Note that these statements may contain C<DROP TABLE> or
2426 C<DROP COLUMN> statements that can potentially destroy data.
2428 The file names are created using the C<ddl_filename> method below, please
2429 override this method in your schema if you would like a different file
2430 name format. For the ALTER file, the same format is used, replacing
2431 $version in the name with "$preversion-$version".
2433 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2434 The most common value for this would be C<< { add_drop_table => 1 } >>
2435 to have the SQL produced include a C<DROP TABLE> statement for each table
2436 created. For quoting purposes supply C<quote_table_names> and
2437 C<quote_field_names>.
2439 If no arguments are passed, then the following default values are assumed:
2443 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2445 =item version - $schema->schema_version
2447 =item directory - './'
2449 =item preversion - <none>
2453 By default, C<\%sqlt_args> will have
2455 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2457 merged with the hash passed in. To disable any of those features, pass in a
2458 hashref like the following
2460 { ignore_constraint_names => 0, # ... other options }
2463 WARNING: You are strongly advised to check all SQL files created, before applying
2468 sub create_ddl_dir {
2469 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2472 carp "No directory given, using ./\n";
2477 make_path ("$dir") # make_path does not like objects (i.e. Path::Class::Dir)
2479 $self->throw_exception(
2480 "Failed to create '$dir': " . ($! || $@ || 'error unknow')
2484 $self->throw_exception ("Directory '$dir' does not exist\n") unless(-d $dir);
2486 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2487 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2489 my $schema_version = $schema->schema_version || '1.x';
2490 $version ||= $schema_version;
2493 add_drop_table => 1,
2494 ignore_constraint_names => 1,
2495 ignore_index_names => 1,
2499 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy')) {
2500 $self->throw_exception("Can't create a ddl file without " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2503 my $sqlt = SQL::Translator->new( $sqltargs );
2505 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2506 my $sqlt_schema = $sqlt->translate({ data => $schema })
2507 or $self->throw_exception ($sqlt->error);
2509 foreach my $db (@$databases) {
2511 $sqlt->{schema} = $sqlt_schema;
2512 $sqlt->producer($db);
2515 my $filename = $schema->ddl_filename($db, $version, $dir);
2516 if (-e $filename && ($version eq $schema_version )) {
2517 # if we are dumping the current version, overwrite the DDL
2518 carp "Overwriting existing DDL file - $filename";
2522 my $output = $sqlt->translate;
2524 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2527 if(!open($file, ">$filename")) {
2528 $self->throw_exception("Can't open $filename for writing ($!)");
2531 print $file $output;
2534 next unless ($preversion);
2536 require SQL::Translator::Diff;
2538 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2539 if(!-e $prefilename) {
2540 carp("No previous schema file found ($prefilename)");
2544 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2546 carp("Overwriting existing diff file - $difffile");
2552 my $t = SQL::Translator->new($sqltargs);
2557 or $self->throw_exception ($t->error);
2559 my $out = $t->translate( $prefilename )
2560 or $self->throw_exception ($t->error);
2562 $source_schema = $t->schema;
2564 $source_schema->name( $prefilename )
2565 unless ( $source_schema->name );
2568 # The "new" style of producers have sane normalization and can support
2569 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2570 # And we have to diff parsed SQL against parsed SQL.
2571 my $dest_schema = $sqlt_schema;
2573 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2574 my $t = SQL::Translator->new($sqltargs);
2579 or $self->throw_exception ($t->error);
2581 my $out = $t->translate( $filename )
2582 or $self->throw_exception ($t->error);
2584 $dest_schema = $t->schema;
2586 $dest_schema->name( $filename )
2587 unless $dest_schema->name;
2590 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2594 if(!open $file, ">$difffile") {
2595 $self->throw_exception("Can't write to $difffile ($!)");
2603 =head2 deployment_statements
2607 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2611 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2613 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2614 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2616 C<$directory> is used to return statements from files in a previously created
2617 L</create_ddl_dir> directory and is optional. The filenames are constructed
2618 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2620 If no C<$directory> is specified then the statements are constructed on the
2621 fly using L<SQL::Translator> and C<$version> is ignored.
2623 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2627 sub deployment_statements {
2628 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2629 $type ||= $self->sqlt_type;
2630 $version ||= $schema->schema_version || '1.x';
2632 my $filename = $schema->ddl_filename($type, $version, $dir);
2636 open($file, "<$filename")
2637 or $self->throw_exception("Can't open $filename ($!)");
2640 return join('', @rows);
2643 unless (DBIx::Class::Optional::Dependencies->req_ok_for ('deploy') ) {
2644 $self->throw_exception("Can't deploy without a ddl_dir or " . DBIx::Class::Optional::Dependencies->req_missing_for ('deploy') );
2647 # sources needs to be a parser arg, but for simplicty allow at top level
2649 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2650 if exists $sqltargs->{sources};
2652 my $tr = SQL::Translator->new(
2653 producer => "SQL::Translator::Producer::${type}",
2655 parser => 'SQL::Translator::Parser::DBIx::Class',
2661 @ret = $tr->translate;
2664 $ret[0] = $tr->translate;
2667 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
2668 unless (@ret && defined $ret[0]);
2670 return wantarray ? @ret : $ret[0];
2674 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2677 return if($line =~ /^--/);
2679 # next if($line =~ /^DROP/m);
2680 return if($line =~ /^BEGIN TRANSACTION/m);
2681 return if($line =~ /^COMMIT/m);
2682 return if $line =~ /^\s+$/; # skip whitespace only
2683 $self->_query_start($line);
2685 # do a dbh_do cycle here, as we need some error checking in
2686 # place (even though we will ignore errors)
2687 $self->dbh_do (sub { $_[1]->do($line) });
2689 carp qq{$_ (running "${line}")};
2691 $self->_query_end($line);
2693 my @statements = $schema->deployment_statements($type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2694 if (@statements > 1) {
2695 foreach my $statement (@statements) {
2696 $deploy->( $statement );
2699 elsif (@statements == 1) {
2700 foreach my $line ( split(";\n", $statements[0])) {
2706 =head2 datetime_parser
2708 Returns the datetime parser class
2712 sub datetime_parser {
2714 return $self->{datetime_parser} ||= do {
2715 $self->build_datetime_parser(@_);
2719 =head2 datetime_parser_type
2721 Defines (returns) the datetime parser class - currently hardwired to
2722 L<DateTime::Format::MySQL>
2726 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2728 =head2 build_datetime_parser
2730 See L</datetime_parser>
2734 sub build_datetime_parser {
2736 my $type = $self->datetime_parser_type(@_);
2737 $self->ensure_class_loaded ($type);
2742 =head2 is_replicating
2744 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2745 replicate from a master database. Default is undef, which is the result
2746 returned by databases that don't support replication.
2750 sub is_replicating {
2755 =head2 lag_behind_master
2757 Returns a number that represents a certain amount of lag behind a master db
2758 when a given storage is replicating. The number is database dependent, but
2759 starts at zero and increases with the amount of lag. Default in undef
2763 sub lag_behind_master {
2767 =head2 relname_to_table_alias
2771 =item Arguments: $relname, $join_count
2775 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
2778 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
2779 way these aliases are named.
2781 The default behavior is C<< "$relname_$join_count" if $join_count > 1 >>,
2782 otherwise C<"$relname">.
2786 sub relname_to_table_alias {
2787 my ($self, $relname, $join_count) = @_;
2789 my $alias = ($join_count && $join_count > 1 ?
2790 join('_', $relname, $join_count) : $relname);
2795 # The size in bytes to use for DBI's ->bind_param_inout, this is the generic
2796 # version and it may be necessary to amend or override it for a specific storage
2797 # if such binds are necessary.
2798 sub _max_column_bytesize {
2799 my ($self, $source, $col) = @_;
2801 my $inf = $source->column_info($col);
2802 return $inf->{_max_bytesize} ||= do {
2806 if (my $data_type = $inf->{data_type}) {
2807 $data_type = lc($data_type);
2809 # String/sized-binary types
2810 if ($data_type =~ /^(?:l?(?:var)?char(?:acter)?(?:\s*varying)?
2811 |(?:var)?binary(?:\s*varying)?|raw)\b/x
2813 $max_size = $inf->{size};
2815 # Other charset/unicode types, assume scale of 4
2816 elsif ($data_type =~ /^(?:national\s*character(?:\s*varying)?|nchar
2820 $max_size = $inf->{size} * 4 if $inf->{size};
2823 elsif ($self->_is_lob_type($data_type)) {
2824 # default to longreadlen
2827 $max_size = 100; # for all other (numeric?) datatypes
2831 $max_size ||= $self->_get_dbh->{LongReadLen} || 8000;
2835 # Determine if a data_type is some type of BLOB
2837 my ($self, $data_type) = @_;
2838 $data_type && ($data_type =~ /(?:lob|bfile|text|image|bytea|memo)/i
2839 || $data_type =~ /^long(?:\s*(?:raw|bit\s*varying|varbit|binary
2840 |varchar|character\s*varying|nvarchar
2841 |national\s*character\s*varying))?$/xi);
2848 =head2 DBIx::Class and AutoCommit
2850 DBIx::Class can do some wonderful magic with handling exceptions,
2851 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2852 (the default) combined with C<txn_do> for transaction support.
2854 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2855 in an assumed transaction between commits, and you're telling us you'd
2856 like to manage that manually. A lot of the magic protections offered by
2857 this module will go away. We can't protect you from exceptions due to database
2858 disconnects because we don't know anything about how to restart your
2859 transactions. You're on your own for handling all sorts of exceptional
2860 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2866 Matt S. Trout <mst@shadowcatsystems.co.uk>
2868 Andy Grundman <andy@hybridized.org>
2872 You may distribute this code under the same terms as Perl itself.