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/;
12 use DBIx::Class::Storage::DBI::Cursor;
13 use DBIx::Class::Storage::Statistics;
16 use Data::Dumper::Concise();
19 # what version of sqlt do we require if deploy() without a ddl_dir is invoked
20 # when changing also adjust the corresponding author_require in Makefile.PL
21 my $minimum_sqlt_version = '0.11002';
24 __PACKAGE__->mk_group_accessors('simple' =>
25 qw/_connect_info _dbi_connect_info _dbh _sql_maker _sql_maker_opts _conn_pid
26 _conn_tid transaction_depth _dbh_autocommit _driver_determined savepoints/
29 # the values for these accessors are picked out (and deleted) from
30 # the attribute hashref passed to connect_info
31 my @storage_options = qw/
32 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
33 disable_sth_caching unsafe auto_savepoint
35 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
38 # default cursor class, overridable in connect_info attributes
39 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
41 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class/);
42 __PACKAGE__->sql_maker_class('DBIx::Class::SQLAHacks');
45 # Each of these methods need _determine_driver called before itself
46 # in order to function reliably. This is a purely DRY optimization
47 my @rdbms_specific_methods = qw/
61 for my $meth (@rdbms_specific_methods) {
63 my $orig = __PACKAGE__->can ($meth)
67 no warnings qw/redefine/;
68 *{__PACKAGE__ ."::$meth"} = Sub::Name::subname $meth => sub {
69 if (not $_[0]->_driver_determined) {
70 $_[0]->_determine_driver;
71 goto $_[0]->can($meth);
80 DBIx::Class::Storage::DBI - DBI storage handler
84 my $schema = MySchema->connect('dbi:SQLite:my.db');
86 $schema->storage->debug(1);
88 my @stuff = $schema->storage->dbh_do(
90 my ($storage, $dbh, @args) = @_;
91 $dbh->do("DROP TABLE authors");
96 $schema->resultset('Book')->search({
97 written_on => $schema->storage->datetime_parser(DateTime->now)
102 This class represents the connection to an RDBMS via L<DBI>. See
103 L<DBIx::Class::Storage> for general information. This pod only
104 documents DBI-specific methods and behaviors.
111 my $new = shift->next::method(@_);
113 $new->transaction_depth(0);
114 $new->_sql_maker_opts({});
115 $new->{savepoints} = [];
116 $new->{_in_dbh_do} = 0;
117 $new->{_dbh_gen} = 0;
124 This method is normally called by L<DBIx::Class::Schema/connection>, which
125 encapsulates its argument list in an arrayref before passing them here.
127 The argument list may contain:
133 The same 4-element argument set one would normally pass to
134 L<DBI/connect>, optionally followed by
135 L<extra attributes|/DBIx::Class specific connection attributes>
136 recognized by DBIx::Class:
138 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
142 A single code reference which returns a connected
143 L<DBI database handle|DBI/connect> optionally followed by
144 L<extra attributes|/DBIx::Class specific connection attributes> recognized
147 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
151 A single hashref with all the attributes and the dsn/user/password
154 $connect_info_args = [{
162 $connect_info_args = [{
163 dbh_maker => sub { DBI->connect (...) },
168 This is particularly useful for L<Catalyst> based applications, allowing the
169 following config (L<Config::General> style):
174 dsn dbi:mysql:database=test
181 The C<dsn>/C<user>/C<password> combination can be substituted by the
182 C<dbh_maker> key whose value is a coderef that returns a connected
183 L<DBI database handle|DBI/connect>
187 Please note that the L<DBI> docs recommend that you always explicitly
188 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
189 recommends that it be set to I<1>, and that you perform transactions
190 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
191 to I<1> if you do not do explicitly set it to zero. This is the default
192 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
194 =head3 DBIx::Class specific connection attributes
196 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
197 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
198 the following connection options. These options can be mixed in with your other
199 L<DBI> connection attributes, or placed in a seperate hashref
200 (C<\%extra_attributes>) as shown above.
202 Every time C<connect_info> is invoked, any previous settings for
203 these options will be cleared before setting the new ones, regardless of
204 whether any options are specified in the new C<connect_info>.
211 Specifies things to do immediately after connecting or re-connecting to
212 the database. Its value may contain:
218 This contains one SQL statement to execute.
220 =item an array reference
222 This contains SQL statements to execute in order. Each element contains
223 a string or a code reference that returns a string.
225 =item a code reference
227 This contains some code to execute. Unlike code references within an
228 array reference, its return value is ignored.
232 =item on_disconnect_do
234 Takes arguments in the same form as L</on_connect_do> and executes them
235 immediately before disconnecting from the database.
237 Note, this only runs if you explicitly call L</disconnect> on the
240 =item on_connect_call
242 A more generalized form of L</on_connect_do> that calls the specified
243 C<connect_call_METHOD> methods in your storage driver.
245 on_connect_do => 'select 1'
249 on_connect_call => [ [ do_sql => 'select 1' ] ]
251 Its values may contain:
257 Will call the C<connect_call_METHOD> method.
259 =item a code reference
261 Will execute C<< $code->($storage) >>
263 =item an array reference
265 Each value can be a method name or code reference.
267 =item an array of arrays
269 For each array, the first item is taken to be the C<connect_call_> method name
270 or code reference, and the rest are parameters to it.
274 Some predefined storage methods you may use:
280 Executes a SQL string or a code reference that returns a SQL string. This is
281 what L</on_connect_do> and L</on_disconnect_do> use.
289 Will execute the scalar as SQL.
293 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
294 attributes hashref and bind values.
296 =item a code reference
298 Will execute C<< $code->($storage) >> and execute the return array refs as
305 Execute any statements necessary to initialize the database session to return
306 and accept datetime/timestamp values used with
307 L<DBIx::Class::InflateColumn::DateTime>.
309 Only necessary for some databases, see your specific storage driver for
310 implementation details.
314 =item on_disconnect_call
316 Takes arguments in the same form as L</on_connect_call> and executes them
317 immediately before disconnecting from the database.
319 Calls the C<disconnect_call_METHOD> methods as opposed to the
320 C<connect_call_METHOD> methods called by L</on_connect_call>.
322 Note, this only runs if you explicitly call L</disconnect> on the
325 =item disable_sth_caching
327 If set to a true value, this option will disable the caching of
328 statement handles via L<DBI/prepare_cached>.
332 Sets the limit dialect. This is useful for JDBC-bridge among others
333 where the remote SQL-dialect cannot be determined by the name of the
334 driver alone. See also L<SQL::Abstract::Limit>.
338 Specifies what characters to use to quote table and column names. If
339 you use this you will want to specify L</name_sep> as well.
341 C<quote_char> expects either a single character, in which case is it
342 is placed on either side of the table/column name, or an arrayref of length
343 2 in which case the table/column name is placed between the elements.
345 For example under MySQL you should use C<< quote_char => '`' >>, and for
346 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
350 This only needs to be used in conjunction with C<quote_char>, and is used to
351 specify the charecter that seperates elements (schemas, tables, columns) from
352 each other. In most cases this is simply a C<.>.
354 The consequences of not supplying this value is that L<SQL::Abstract>
355 will assume DBIx::Class' uses of aliases to be complete column
356 names. The output will look like I<"me.name"> when it should actually
361 This Storage driver normally installs its own C<HandleError>, sets
362 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
363 all database handles, including those supplied by a coderef. It does this
364 so that it can have consistent and useful error behavior.
366 If you set this option to a true value, Storage will not do its usual
367 modifications to the database handle's attributes, and instead relies on
368 the settings in your connect_info DBI options (or the values you set in
369 your connection coderef, in the case that you are connecting via coderef).
371 Note that your custom settings can cause Storage to malfunction,
372 especially if you set a C<HandleError> handler that suppresses exceptions
373 and/or disable C<RaiseError>.
377 If this option is true, L<DBIx::Class> will use savepoints when nesting
378 transactions, making it possible to recover from failure in the inner
379 transaction without having to abort all outer transactions.
383 Use this argument to supply a cursor class other than the default
384 L<DBIx::Class::Storage::DBI::Cursor>.
388 Some real-life examples of arguments to L</connect_info> and
389 L<DBIx::Class::Schema/connect>
391 # Simple SQLite connection
392 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
395 ->connect_info([ sub { DBI->connect(...) } ]);
397 # Connect via subref in hashref
399 dbh_maker => sub { DBI->connect(...) },
400 on_connect_do => 'alter session ...',
403 # A bit more complicated
410 { quote_char => q{"}, name_sep => q{.} },
414 # Equivalent to the previous example
420 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
424 # Same, but with hashref as argument
425 # See parse_connect_info for explanation
428 dsn => 'dbi:Pg:dbname=foo',
430 password => 'my_pg_password',
437 # Subref + DBIx::Class-specific connection options
440 sub { DBI->connect(...) },
444 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
445 disable_sth_caching => 1,
455 my ($self, $info) = @_;
457 return $self->_connect_info if !$info;
459 $self->_connect_info($info); # copy for _connect_info
461 $info = $self->_normalize_connect_info($info)
462 if ref $info eq 'ARRAY';
464 for my $storage_opt (keys %{ $info->{storage_options} }) {
465 my $value = $info->{storage_options}{$storage_opt};
467 $self->$storage_opt($value);
470 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
471 # the new set of options
472 $self->_sql_maker(undef);
473 $self->_sql_maker_opts({});
475 for my $sql_maker_opt (keys %{ $info->{sql_maker_options} }) {
476 my $value = $info->{sql_maker_options}{$sql_maker_opt};
478 $self->_sql_maker_opts->{$sql_maker_opt} = $value;
482 %{ $self->_default_dbi_connect_attributes || {} },
483 %{ $info->{attributes} || {} },
486 my @args = @{ $info->{arguments} };
488 $self->_dbi_connect_info([@args,
489 %attrs && !(ref $args[0] eq 'CODE') ? \%attrs : ()]);
491 return $self->_connect_info;
494 sub _normalize_connect_info {
495 my ($self, $info_arg) = @_;
498 my @args = @$info_arg; # take a shallow copy for further mutilation
500 # combine/pre-parse arguments depending on invocation style
503 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
504 %attrs = %{ $args[1] || {} };
507 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
508 %attrs = %{$args[0]};
510 if (my $code = delete $attrs{dbh_maker}) {
513 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
516 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
517 . "to the result of 'dbh_maker'",
519 join (', ', map { "'$_'" } (@ignored) ),
524 @args = delete @attrs{qw/dsn user password/};
527 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
529 % { $args[3] || {} },
530 % { $args[4] || {} },
532 @args = @args[0,1,2];
535 $info{arguments} = \@args;
537 my @storage_opts = grep exists $attrs{$_},
538 @storage_options, 'cursor_class';
540 @{ $info{storage_options} }{@storage_opts} =
541 delete @attrs{@storage_opts} if @storage_opts;
543 my @sql_maker_opts = grep exists $attrs{$_},
544 qw/limit_dialect quote_char name_sep/;
546 @{ $info{sql_maker_options} }{@sql_maker_opts} =
547 delete @attrs{@sql_maker_opts} if @sql_maker_opts;
549 $info{attributes} = \%attrs if %attrs;
554 sub _default_dbi_connect_attributes {
564 This method is deprecated in favour of setting via L</connect_info>.
568 =head2 on_disconnect_do
570 This method is deprecated in favour of setting via L</connect_info>.
574 sub _parse_connect_do {
575 my ($self, $type) = @_;
577 my $val = $self->$type;
578 return () if not defined $val;
583 push @res, [ 'do_sql', $val ];
584 } elsif (ref($val) eq 'CODE') {
586 } elsif (ref($val) eq 'ARRAY') {
587 push @res, map { [ 'do_sql', $_ ] } @$val;
589 $self->throw_exception("Invalid type for $type: ".ref($val));
597 Arguments: ($subref | $method_name), @extra_coderef_args?
599 Execute the given $subref or $method_name using the new exception-based
600 connection management.
602 The first two arguments will be the storage object that C<dbh_do> was called
603 on and a database handle to use. Any additional arguments will be passed
604 verbatim to the called subref as arguments 2 and onwards.
606 Using this (instead of $self->_dbh or $self->dbh) ensures correct
607 exception handling and reconnection (or failover in future subclasses).
609 Your subref should have no side-effects outside of the database, as
610 there is the potential for your subref to be partially double-executed
611 if the database connection was stale/dysfunctional.
615 my @stuff = $schema->storage->dbh_do(
617 my ($storage, $dbh, @cols) = @_;
618 my $cols = join(q{, }, @cols);
619 $dbh->selectrow_array("SELECT $cols FROM foo");
630 my $dbh = $self->_get_dbh;
632 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
633 || $self->{transaction_depth};
635 local $self->{_in_dbh_do} = 1;
638 my $want_array = wantarray;
643 @result = $self->$code($dbh, @_);
645 elsif(defined $want_array) {
646 $result[0] = $self->$code($dbh, @_);
649 $self->$code($dbh, @_);
653 # ->connected might unset $@ - copy
655 if(!$exception) { return $want_array ? @result : $result[0] }
657 $self->throw_exception($exception) if $self->connected;
659 # We were not connected - reconnect and retry, but let any
660 # exception fall right through this time
661 carp "Retrying $code after catching disconnected exception: $exception"
662 if $ENV{DBIC_DBIRETRY_DEBUG};
663 $self->_populate_dbh;
664 $self->$code($self->_dbh, @_);
667 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
668 # It also informs dbh_do to bypass itself while under the direction of txn_do,
669 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
674 ref $coderef eq 'CODE' or $self->throw_exception
675 ('$coderef must be a CODE reference');
677 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
679 local $self->{_in_dbh_do} = 1;
682 my $want_array = wantarray;
691 @result = $coderef->(@_);
693 elsif(defined $want_array) {
694 $result[0] = $coderef->(@_);
702 # ->connected might unset $@ - copy
704 if(!$exception) { return $want_array ? @result : $result[0] }
706 if($tried++ || $self->connected) {
707 eval { $self->txn_rollback };
708 my $rollback_exception = $@;
709 if($rollback_exception) {
710 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
711 $self->throw_exception($exception) # propagate nested rollback
712 if $rollback_exception =~ /$exception_class/;
714 $self->throw_exception(
715 "Transaction aborted: ${exception}. "
716 . "Rollback failed: ${rollback_exception}"
719 $self->throw_exception($exception)
722 # We were not connected, and was first try - reconnect and retry
724 carp "Retrying $coderef after catching disconnected exception: $exception"
725 if $ENV{DBIC_DBIRETRY_DEBUG};
726 $self->_populate_dbh;
732 Our C<disconnect> method also performs a rollback first if the
733 database is not in C<AutoCommit> mode.
743 push @actions, ( $self->on_disconnect_call || () );
744 push @actions, $self->_parse_connect_do ('on_disconnect_do');
746 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
748 $self->_dbh_rollback unless $self->_dbh_autocommit;
750 $self->_dbh->disconnect;
756 =head2 with_deferred_fk_checks
760 =item Arguments: C<$coderef>
762 =item Return Value: The return value of $coderef
766 Storage specific method to run the code ref with FK checks deferred or
767 in MySQL's case disabled entirely.
771 # Storage subclasses should override this
772 sub with_deferred_fk_checks {
773 my ($self, $sub) = @_;
781 =item Arguments: none
783 =item Return Value: 1|0
787 Verifies that the the current database handle is active and ready to execute
788 an SQL statement (i.e. the connection did not get stale, server is still
789 answering, etc.) This method is used internally by L</dbh>.
795 return 0 unless $self->_seems_connected;
798 local $self->_dbh->{RaiseError} = 1;
803 sub _seems_connected {
806 my $dbh = $self->_dbh
809 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
816 return 0 if !$self->_dbh;
819 return $dbh->FETCH('Active');
825 my $dbh = $self->_dbh or return 0;
830 # handle pid changes correctly
831 # NOTE: assumes $self->_dbh is a valid $dbh
835 return if defined $self->_conn_pid && $self->_conn_pid == $$;
837 $self->_dbh->{InactiveDestroy} = 1;
844 sub ensure_connected {
847 unless ($self->connected) {
848 $self->_populate_dbh;
854 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
855 is guaranteed to be healthy by implicitly calling L</connected>, and if
856 necessary performing a reconnection before returning. Keep in mind that this
857 is very B<expensive> on some database engines. Consider using L<dbh_do>
865 if (not $self->_dbh) {
866 $self->_populate_dbh;
868 $self->ensure_connected;
873 # this is the internal "get dbh or connect (don't check)" method
876 $self->_verify_pid if $self->_dbh;
877 $self->_populate_dbh unless $self->_dbh;
881 sub _sql_maker_args {
886 array_datatypes => 1,
887 limit_dialect => $self->_get_dbh,
888 %{$self->_sql_maker_opts}
894 unless ($self->_sql_maker) {
895 my $sql_maker_class = $self->sql_maker_class;
896 $self->ensure_class_loaded ($sql_maker_class);
897 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
899 return $self->_sql_maker;
902 # nothing to do by default
909 my @info = @{$self->_dbi_connect_info || []};
910 $self->_dbh(undef); # in case ->connected failed we might get sent here
911 $self->_dbh($self->_connect(@info));
913 $self->_conn_pid($$);
914 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
916 $self->_determine_driver;
918 # Always set the transaction depth on connect, since
919 # there is no transaction in progress by definition
920 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
922 $self->_run_connection_actions unless $self->{_in_determine_driver};
925 sub _run_connection_actions {
929 push @actions, ( $self->on_connect_call || () );
930 push @actions, $self->_parse_connect_do ('on_connect_do');
932 $self->_do_connection_actions(connect_call_ => $_) for @actions;
935 sub _determine_driver {
938 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
939 my $started_connected = 0;
940 local $self->{_in_determine_driver} = 1;
942 if (ref($self) eq __PACKAGE__) {
944 if ($self->_dbh) { # we are connected
945 $driver = $self->_dbh->{Driver}{Name};
946 $started_connected = 1;
948 # if connect_info is a CODEREF, we have no choice but to connect
949 if (ref $self->_dbi_connect_info->[0] &&
950 Scalar::Util::reftype($self->_dbi_connect_info->[0]) eq 'CODE') {
951 $self->_populate_dbh;
952 $driver = $self->_dbh->{Driver}{Name};
955 # try to use dsn to not require being connected, the driver may still
956 # force a connection in _rebless to determine version
957 ($driver) = $self->_dbi_connect_info->[0] =~ /dbi:([^:]+):/i;
961 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
962 if ($self->load_optional_class($storage_class)) {
963 mro::set_mro($storage_class, 'c3');
964 bless $self, $storage_class;
969 $self->_driver_determined(1);
971 $self->_init; # run driver-specific initializations
973 $self->_run_connection_actions
974 if !$started_connected && defined $self->_dbh;
978 sub _do_connection_actions {
980 my $method_prefix = shift;
983 if (not ref($call)) {
984 my $method = $method_prefix . $call;
986 } elsif (ref($call) eq 'CODE') {
988 } elsif (ref($call) eq 'ARRAY') {
989 if (ref($call->[0]) ne 'ARRAY') {
990 $self->_do_connection_actions($method_prefix, $_) for @$call;
992 $self->_do_connection_actions($method_prefix, @$_) for @$call;
995 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
1001 sub connect_call_do_sql {
1003 $self->_do_query(@_);
1006 sub disconnect_call_do_sql {
1008 $self->_do_query(@_);
1011 # override in db-specific backend when necessary
1012 sub connect_call_datetime_setup { 1 }
1015 my ($self, $action) = @_;
1017 if (ref $action eq 'CODE') {
1018 $action = $action->($self);
1019 $self->_do_query($_) foreach @$action;
1022 # Most debuggers expect ($sql, @bind), so we need to exclude
1023 # the attribute hash which is the second argument to $dbh->do
1024 # furthermore the bind values are usually to be presented
1025 # as named arrayref pairs, so wrap those here too
1026 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
1027 my $sql = shift @do_args;
1028 my $attrs = shift @do_args;
1029 my @bind = map { [ undef, $_ ] } @do_args;
1031 $self->_query_start($sql, @bind);
1032 $self->_get_dbh->do($sql, $attrs, @do_args);
1033 $self->_query_end($sql, @bind);
1040 my ($self, @info) = @_;
1042 $self->throw_exception("You failed to provide any connection info")
1045 my ($old_connect_via, $dbh);
1047 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
1048 $old_connect_via = $DBI::connect_via;
1049 $DBI::connect_via = 'connect';
1053 if(ref $info[0] eq 'CODE') {
1054 $dbh = $info[0]->();
1057 $dbh = DBI->connect(@info);
1060 if($dbh && !$self->unsafe) {
1061 my $weak_self = $self;
1062 Scalar::Util::weaken($weak_self);
1063 $dbh->{HandleError} = sub {
1065 $weak_self->throw_exception("DBI Exception: $_[0]");
1068 # the handler may be invoked by something totally out of
1070 croak ("DBI Exception: $_[0]");
1073 $dbh->{ShowErrorStatement} = 1;
1074 $dbh->{RaiseError} = 1;
1075 $dbh->{PrintError} = 0;
1079 $DBI::connect_via = $old_connect_via if $old_connect_via;
1081 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
1084 $self->_dbh_autocommit($dbh->{AutoCommit});
1090 my ($self, $name) = @_;
1092 $name = $self->_svp_generate_name
1093 unless defined $name;
1095 $self->throw_exception ("You can't use savepoints outside a transaction")
1096 if $self->{transaction_depth} == 0;
1098 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1099 unless $self->can('_svp_begin');
1101 push @{ $self->{savepoints} }, $name;
1103 $self->debugobj->svp_begin($name) if $self->debug;
1105 return $self->_svp_begin($name);
1109 my ($self, $name) = @_;
1111 $self->throw_exception ("You can't use savepoints outside a transaction")
1112 if $self->{transaction_depth} == 0;
1114 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1115 unless $self->can('_svp_release');
1117 if (defined $name) {
1118 $self->throw_exception ("Savepoint '$name' does not exist")
1119 unless grep { $_ eq $name } @{ $self->{savepoints} };
1121 # Dig through the stack until we find the one we are releasing. This keeps
1122 # the stack up to date.
1125 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1127 $name = pop @{ $self->{savepoints} };
1130 $self->debugobj->svp_release($name) if $self->debug;
1132 return $self->_svp_release($name);
1136 my ($self, $name) = @_;
1138 $self->throw_exception ("You can't use savepoints outside a transaction")
1139 if $self->{transaction_depth} == 0;
1141 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1142 unless $self->can('_svp_rollback');
1144 if (defined $name) {
1145 # If they passed us a name, verify that it exists in the stack
1146 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1147 $self->throw_exception("Savepoint '$name' does not exist!");
1150 # Dig through the stack until we find the one we are releasing. This keeps
1151 # the stack up to date.
1152 while(my $s = pop(@{ $self->{savepoints} })) {
1153 last if($s eq $name);
1155 # Add the savepoint back to the stack, as a rollback doesn't remove the
1156 # named savepoint, only everything after it.
1157 push(@{ $self->{savepoints} }, $name);
1159 # We'll assume they want to rollback to the last savepoint
1160 $name = $self->{savepoints}->[-1];
1163 $self->debugobj->svp_rollback($name) if $self->debug;
1165 return $self->_svp_rollback($name);
1168 sub _svp_generate_name {
1171 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1177 # this means we have not yet connected and do not know the AC status
1178 # (e.g. coderef $dbh)
1179 $self->ensure_connected if (! defined $self->_dbh_autocommit);
1181 if($self->{transaction_depth} == 0) {
1182 $self->debugobj->txn_begin()
1184 $self->_dbh_begin_work;
1186 elsif ($self->auto_savepoint) {
1189 $self->{transaction_depth}++;
1192 sub _dbh_begin_work {
1195 # if the user is utilizing txn_do - good for him, otherwise we need to
1196 # ensure that the $dbh is healthy on BEGIN.
1197 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1198 # will be replaced by a failure of begin_work itself (which will be
1199 # then retried on reconnect)
1200 if ($self->{_in_dbh_do}) {
1201 $self->_dbh->begin_work;
1203 $self->dbh_do(sub { $_[1]->begin_work });
1209 if ($self->{transaction_depth} == 1) {
1210 $self->debugobj->txn_commit()
1213 $self->{transaction_depth} = 0
1214 if $self->_dbh_autocommit;
1216 elsif($self->{transaction_depth} > 1) {
1217 $self->{transaction_depth}--;
1219 if $self->auto_savepoint;
1225 my $dbh = $self->_dbh
1226 or $self->throw_exception('cannot COMMIT on a disconnected handle');
1232 my $dbh = $self->_dbh;
1234 if ($self->{transaction_depth} == 1) {
1235 $self->debugobj->txn_rollback()
1237 $self->{transaction_depth} = 0
1238 if $self->_dbh_autocommit;
1239 $self->_dbh_rollback;
1241 elsif($self->{transaction_depth} > 1) {
1242 $self->{transaction_depth}--;
1243 if ($self->auto_savepoint) {
1244 $self->svp_rollback;
1249 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1254 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1255 $error =~ /$exception_class/ and $self->throw_exception($error);
1256 # ensure that a failed rollback resets the transaction depth
1257 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1258 $self->throw_exception($error);
1264 my $dbh = $self->_dbh
1265 or $self->throw_exception('cannot ROLLBACK on a disconnected handle');
1269 # This used to be the top-half of _execute. It was split out to make it
1270 # easier to override in NoBindVars without duping the rest. It takes up
1271 # all of _execute's args, and emits $sql, @bind.
1272 sub _prep_for_execute {
1273 my ($self, $op, $extra_bind, $ident, $args) = @_;
1275 if( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1276 $ident = $ident->from();
1279 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1282 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1284 return ($sql, \@bind);
1288 sub _fix_bind_params {
1289 my ($self, @bind) = @_;
1291 ### Turn @bind from something like this:
1292 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1294 ### ( "'1'", "'1'", "'3'" )
1297 if ( defined( $_ && $_->[1] ) ) {
1298 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1305 my ( $self, $sql, @bind ) = @_;
1307 if ( $self->debug ) {
1308 @bind = $self->_fix_bind_params(@bind);
1310 $self->debugobj->query_start( $sql, @bind );
1315 my ( $self, $sql, @bind ) = @_;
1317 if ( $self->debug ) {
1318 @bind = $self->_fix_bind_params(@bind);
1319 $self->debugobj->query_end( $sql, @bind );
1324 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1326 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1328 $self->_query_start( $sql, @$bind );
1330 my $sth = $self->sth($sql,$op);
1332 my $placeholder_index = 1;
1334 foreach my $bound (@$bind) {
1335 my $attributes = {};
1336 my($column_name, @data) = @$bound;
1338 if ($bind_attributes) {
1339 $attributes = $bind_attributes->{$column_name}
1340 if defined $bind_attributes->{$column_name};
1343 foreach my $data (@data) {
1344 my $ref = ref $data;
1345 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1347 $sth->bind_param($placeholder_index, $data, $attributes);
1348 $placeholder_index++;
1352 # Can this fail without throwing an exception anyways???
1353 my $rv = $sth->execute();
1354 $self->throw_exception($sth->errstr) if !$rv;
1356 $self->_query_end( $sql, @$bind );
1358 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1363 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1367 my ($self, $source, $to_insert) = @_;
1369 my $ident = $source->from;
1370 my $bind_attributes = $self->source_bind_attributes($source);
1372 my $updated_cols = {};
1374 foreach my $col ( $source->columns ) {
1375 if ( !defined $to_insert->{$col} ) {
1376 my $col_info = $source->column_info($col);
1378 if ( $col_info->{auto_nextval} ) {
1379 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1381 $col_info->{sequence} ||
1382 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source)
1388 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
1390 return $updated_cols;
1393 ## Currently it is assumed that all values passed will be "normal", i.e. not
1394 ## scalar refs, or at least, all the same type as the first set, the statement is
1395 ## only prepped once.
1397 my ($self, $source, $cols, $data) = @_;
1400 @colvalues{@$cols} = (0..$#$cols);
1402 for my $i (0..$#$cols) {
1403 my $first_val = $data->[0][$i];
1404 next unless ref $first_val eq 'SCALAR';
1406 $colvalues{ $cols->[$i] } = $first_val;
1409 # check for bad data and stringify stringifiable objects
1410 my $bad_slice = sub {
1411 my ($msg, $col_idx, $slice_idx) = @_;
1412 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
1416 local $Data::Dumper::Maxdepth = 1; # don't dump objects, if any
1417 Data::Dumper::Concise::Dumper({
1418 map { $cols->[$_] => $data->[$slice_idx][$_] } (0 .. $#$cols)
1424 for my $datum_idx (0..$#$data) {
1425 my $datum = $data->[$datum_idx];
1427 for my $col_idx (0..$#$cols) {
1428 my $val = $datum->[$col_idx];
1429 my $sqla_bind = $colvalues{ $cols->[$col_idx] };
1430 my $is_literal_sql = (ref $sqla_bind) eq 'SCALAR';
1432 if ($is_literal_sql) {
1434 $bad_slice->('bind found where literal SQL expected', $col_idx, $datum_idx);
1436 elsif ((my $reftype = ref $val) ne 'SCALAR') {
1437 $bad_slice->("$reftype reference found where literal SQL expected",
1438 $col_idx, $datum_idx);
1440 elsif ($$val ne $$sqla_bind){
1441 $bad_slice->("inconsistent literal SQL value, expecting: '$$sqla_bind'",
1442 $col_idx, $datum_idx);
1445 elsif (my $reftype = ref $val) {
1447 if (overload::Method($val, '""')) {
1448 $datum->[$col_idx] = "".$val;
1451 $bad_slice->("$reftype reference found where bind expected",
1452 $col_idx, $datum_idx);
1458 my ($sql, $bind) = $self->_prep_for_execute (
1459 'insert', undef, $source, [\%colvalues]
1463 my $empty_bind = 1 if (not @bind) &&
1464 (grep { ref $_ eq 'SCALAR' } values %colvalues) == @$cols;
1466 if ((not @bind) && (not $empty_bind)) {
1467 $self->throw_exception(
1468 'Cannot insert_bulk without support for placeholders'
1472 # neither _execute_array, nor _execute_inserts_with_no_binds are
1473 # atomic (even if _execute _array is a single call). Thus a safety
1475 my $guard = $self->txn_scope_guard unless $self->{transaction_depth} != 0;
1477 $self->_query_start( $sql, ['__BULK__'] );
1478 my $sth = $self->sth($sql);
1481 # bind_param_array doesn't work if there are no binds
1482 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
1485 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1486 $self->_execute_array( $source, $sth, \@bind, $cols, $data );
1490 $self->_query_end( $sql, ['__BULK__'] );
1493 $guard->commit if $guard;
1495 return (wantarray ? ($rv, $sth, @bind) : $rv);
1498 sub _execute_array {
1499 my ($self, $source, $sth, $bind, $cols, $data, @extra) = @_;
1501 ## This must be an arrayref, else nothing works!
1502 my $tuple_status = [];
1504 ## Get the bind_attributes, if any exist
1505 my $bind_attributes = $self->source_bind_attributes($source);
1507 ## Bind the values and execute
1508 my $placeholder_index = 1;
1510 foreach my $bound (@$bind) {
1512 my $attributes = {};
1513 my ($column_name, $data_index) = @$bound;
1515 if( $bind_attributes ) {
1516 $attributes = $bind_attributes->{$column_name}
1517 if defined $bind_attributes->{$column_name};
1520 my @data = map { $_->[$data_index] } @$data;
1522 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1523 $placeholder_index++;
1527 $self->_dbh_execute_array($sth, $tuple_status, @extra);
1529 my $err = $@ || $sth->errstr;
1531 # Statement must finish even if there was an exception.
1532 eval { $sth->finish };
1533 $err = $@ unless $err;
1537 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1539 $self->throw_exception("Unexpected populate error: $err")
1540 if ($i > $#$tuple_status);
1542 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1543 ($tuple_status->[$i][1] || $err),
1544 Data::Dumper::Concise::Dumper({
1545 map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols)
1552 sub _dbh_execute_array {
1553 my ($self, $sth, $tuple_status, @extra) = @_;
1555 return $sth->execute_array({ArrayTupleStatus => $tuple_status});
1558 sub _dbh_execute_inserts_with_no_binds {
1559 my ($self, $sth, $count) = @_;
1562 my $dbh = $self->_get_dbh;
1563 local $dbh->{RaiseError} = 1;
1564 local $dbh->{PrintError} = 0;
1566 $sth->execute foreach 1..$count;
1570 # Make sure statement is finished even if there was an exception.
1571 eval { $sth->finish };
1572 $exception = $@ unless $exception;
1574 $self->throw_exception($exception) if $exception;
1580 my ($self, $source, @args) = @_;
1582 my $bind_attrs = $self->source_bind_attributes($source);
1584 return $self->_execute('update' => [], $source, $bind_attrs, @args);
1589 my ($self, $source, @args) = @_;
1591 my $bind_attrs = $self->source_bind_attributes($source);
1593 return $self->_execute('delete' => [], $source, $bind_attrs, @args);
1596 # We were sent here because the $rs contains a complex search
1597 # which will require a subquery to select the correct rows
1598 # (i.e. joined or limited resultsets, or non-introspectable conditions)
1600 # Generating a single PK column subquery is trivial and supported
1601 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1602 # Look at _multipk_update_delete()
1603 sub _subq_update_delete {
1605 my ($rs, $op, $values) = @_;
1607 my $rsrc = $rs->result_source;
1609 # quick check if we got a sane rs on our hands
1610 my @pcols = $rsrc->primary_columns;
1612 $self->throw_exception (
1614 "You must declare primary key(s) on source '%s' (via set_primary_key) in order to update or delete complex resultsets",
1615 $rsrc->source_name || $rsrc->from
1620 my $sel = $rs->_resolved_attrs->{select};
1621 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1624 join ("\x00", map { join '.', $rs->{attrs}{alias}, $_ } sort @pcols)
1626 join ("\x00", sort @$sel )
1628 $self->throw_exception (
1629 '_subq_update_delete can not be called on resultsets selecting columns other than the primary keys'
1636 $op eq 'update' ? $values : (),
1637 { $pcols[0] => { -in => $rs->as_query } },
1642 return $self->_multipk_update_delete (@_);
1646 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1647 # resultset update/delete involving subqueries. So by default resort
1648 # to simple (and inefficient) delete_all style per-row opearations,
1649 # while allowing specific storages to override this with a faster
1652 sub _multipk_update_delete {
1653 return shift->_per_row_update_delete (@_);
1656 # This is the default loop used to delete/update rows for multi PK
1657 # resultsets, and used by mysql exclusively (because it can't do anything
1660 # We do not use $row->$op style queries, because resultset update/delete
1661 # is not expected to cascade (this is what delete_all/update_all is for).
1663 # There should be no race conditions as the entire operation is rolled
1666 sub _per_row_update_delete {
1668 my ($rs, $op, $values) = @_;
1670 my $rsrc = $rs->result_source;
1671 my @pcols = $rsrc->primary_columns;
1673 my $guard = $self->txn_scope_guard;
1675 # emulate the return value of $sth->execute for non-selects
1676 my $row_cnt = '0E0';
1678 my $subrs_cur = $rs->cursor;
1679 my @all_pk = $subrs_cur->all;
1680 for my $pks ( @all_pk) {
1683 for my $i (0.. $#pcols) {
1684 $cond->{$pcols[$i]} = $pks->[$i];
1689 $op eq 'update' ? $values : (),
1704 # localization is neccessary as
1705 # 1) there is no infrastructure to pass this around before SQLA2
1706 # 2) _select_args sets it and _prep_for_execute consumes it
1707 my $sql_maker = $self->sql_maker;
1708 local $sql_maker->{_dbic_rs_attrs};
1710 return $self->_execute($self->_select_args(@_));
1713 sub _select_args_to_query {
1716 # localization is neccessary as
1717 # 1) there is no infrastructure to pass this around before SQLA2
1718 # 2) _select_args sets it and _prep_for_execute consumes it
1719 my $sql_maker = $self->sql_maker;
1720 local $sql_maker->{_dbic_rs_attrs};
1722 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $order, $rows, $offset)
1723 # = $self->_select_args($ident, $select, $cond, $attrs);
1724 my ($op, $bind, $ident, $bind_attrs, @args) =
1725 $self->_select_args(@_);
1727 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $order, $rows, $offset ]);
1728 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1729 $prepared_bind ||= [];
1732 ? ($sql, $prepared_bind, $bind_attrs)
1733 : \[ "($sql)", @$prepared_bind ]
1738 my ($self, $ident, $select, $where, $attrs) = @_;
1740 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1742 my $sql_maker = $self->sql_maker;
1743 $sql_maker->{_dbic_rs_attrs} = {
1748 $rs_alias && $alias2source->{$rs_alias}
1749 ? ( _source_handle => $alias2source->{$rs_alias}->handle )
1754 # calculate bind_attrs before possible $ident mangling
1755 my $bind_attrs = {};
1756 for my $alias (keys %$alias2source) {
1757 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
1758 for my $col (keys %$bindtypes) {
1760 my $fqcn = join ('.', $alias, $col);
1761 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
1763 # Unqialified column names are nice, but at the same time can be
1764 # rather ambiguous. What we do here is basically go along with
1765 # the loop, adding an unqualified column slot to $bind_attrs,
1766 # alongside the fully qualified name. As soon as we encounter
1767 # another column by that name (which would imply another table)
1768 # we unset the unqualified slot and never add any info to it
1769 # to avoid erroneous type binding. If this happens the users
1770 # only choice will be to fully qualify his column name
1772 if (exists $bind_attrs->{$col}) {
1773 $bind_attrs->{$col} = {};
1776 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
1783 $attrs->{software_limit}
1785 $sql_maker->_default_limit_syntax eq "GenericSubQ"
1787 $attrs->{software_limit} = 1;
1790 $self->throw_exception("rows attribute must be positive if present")
1791 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1793 # MySQL actually recommends this approach. I cringe.
1794 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1799 # see if we need to tear the prefetch apart otherwise delegate the limiting to the
1800 # storage, unless software limit was requested
1803 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
1805 # limited prefetch with RNO subqueries
1809 $sql_maker->limit_dialect eq 'RowNumberOver'
1811 $attrs->{_prefetch_select}
1813 @{$attrs->{_prefetch_select}}
1817 ( $attrs->{group_by}
1819 @{$attrs->{group_by}}
1821 $attrs->{_prefetch_select}
1823 @{$attrs->{_prefetch_select}}
1826 ($ident, $select, $where, $attrs)
1827 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
1831 ($attrs->{rows} || $attrs->{offset})
1833 $sql_maker->limit_dialect eq 'RowNumberOver'
1835 (ref $ident eq 'ARRAY' && @$ident > 1) # indicates a join
1837 scalar $self->_parse_order_by ($attrs->{order_by})
1839 # the RNO limit dialect above mangles the SQL such that the join gets lost
1840 # wrap a subquery here
1842 push @limit, delete @{$attrs}{qw/rows offset/};
1844 my $subq = $self->_select_args_to_query (
1852 -alias => $attrs->{alias},
1853 -source_handle => $ident->[0]{-source_handle},
1854 $attrs->{alias} => $subq,
1857 # all part of the subquery now
1858 delete @{$attrs}{qw/order_by group_by having/};
1862 elsif (! $attrs->{software_limit} ) {
1863 push @limit, $attrs->{rows}, $attrs->{offset};
1866 # try to simplify the joinmap further (prune unreferenced type-single joins)
1867 $ident = $self->_prune_unused_joins ($ident, $select, $where, $attrs);
1870 # This would be the point to deflate anything found in $where
1871 # (and leave $attrs->{bind} intact). Problem is - inflators historically
1872 # expect a row object. And all we have is a resultsource (it is trivial
1873 # to extract deflator coderefs via $alias2source above).
1875 # I don't see a way forward other than changing the way deflators are
1876 # invoked, and that's just bad...
1880 { $attrs->{$_} ? ( $_ => $attrs->{$_} ) : () }
1881 (qw/order_by group_by having/ )
1884 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $order, @limit);
1887 # Returns a counting SELECT for a simple count
1888 # query. Abstracted so that a storage could override
1889 # this to { count => 'firstcol' } or whatever makes
1890 # sense as a performance optimization
1892 #my ($self, $source, $rs_attrs) = @_;
1893 return { count => '*' };
1896 # Returns a SELECT which will end up in the subselect
1897 # There may or may not be a group_by, as the subquery
1898 # might have been called to accomodate a limit
1900 # Most databases would be happy with whatever ends up
1901 # here, but some choke in various ways.
1903 sub _subq_count_select {
1904 my ($self, $source, $rs_attrs) = @_;
1905 return $rs_attrs->{group_by} if $rs_attrs->{group_by};
1907 my @pcols = map { join '.', $rs_attrs->{alias}, $_ } ($source->primary_columns);
1908 return @pcols ? \@pcols : [ 1 ];
1911 sub source_bind_attributes {
1912 my ($self, $source) = @_;
1914 my $bind_attributes;
1915 foreach my $column ($source->columns) {
1917 my $data_type = $source->column_info($column)->{data_type} || '';
1918 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
1922 return $bind_attributes;
1929 =item Arguments: $ident, $select, $condition, $attrs
1933 Handle a SQL select statement.
1939 my ($ident, $select, $condition, $attrs) = @_;
1940 return $self->cursor_class->new($self, \@_, $attrs);
1945 my ($rv, $sth, @bind) = $self->_select(@_);
1946 my @row = $sth->fetchrow_array;
1947 my @nextrow = $sth->fetchrow_array if @row;
1948 if(@row && @nextrow) {
1949 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
1951 # Need to call finish() to work round broken DBDs
1960 =item Arguments: $sql
1964 Returns a L<DBI> sth (statement handle) for the supplied SQL.
1969 my ($self, $dbh, $sql) = @_;
1971 # 3 is the if_active parameter which avoids active sth re-use
1972 my $sth = $self->disable_sth_caching
1973 ? $dbh->prepare($sql)
1974 : $dbh->prepare_cached($sql, {}, 3);
1976 # XXX You would think RaiseError would make this impossible,
1977 # but apparently that's not true :(
1978 $self->throw_exception($dbh->errstr) if !$sth;
1984 my ($self, $sql) = @_;
1985 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
1988 sub _dbh_columns_info_for {
1989 my ($self, $dbh, $table) = @_;
1991 if ($dbh->can('column_info')) {
1994 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
1995 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
1997 while ( my $info = $sth->fetchrow_hashref() ){
1999 $column_info{data_type} = $info->{TYPE_NAME};
2000 $column_info{size} = $info->{COLUMN_SIZE};
2001 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2002 $column_info{default_value} = $info->{COLUMN_DEF};
2003 my $col_name = $info->{COLUMN_NAME};
2004 $col_name =~ s/^\"(.*)\"$/$1/;
2006 $result{$col_name} = \%column_info;
2009 return \%result if !$@ && scalar keys %result;
2013 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2015 my @columns = @{$sth->{NAME_lc}};
2016 for my $i ( 0 .. $#columns ){
2018 $column_info{data_type} = $sth->{TYPE}->[$i];
2019 $column_info{size} = $sth->{PRECISION}->[$i];
2020 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2022 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2023 $column_info{data_type} = $1;
2024 $column_info{size} = $2;
2027 $result{$columns[$i]} = \%column_info;
2031 foreach my $col (keys %result) {
2032 my $colinfo = $result{$col};
2033 my $type_num = $colinfo->{data_type};
2035 if(defined $type_num && $dbh->can('type_info')) {
2036 my $type_info = $dbh->type_info($type_num);
2037 $type_name = $type_info->{TYPE_NAME} if $type_info;
2038 $colinfo->{data_type} = $type_name if $type_name;
2045 sub columns_info_for {
2046 my ($self, $table) = @_;
2047 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2050 =head2 last_insert_id
2052 Return the row id of the last insert.
2056 sub _dbh_last_insert_id {
2057 # All Storage's need to register their own _dbh_last_insert_id
2058 # the old SQLite-based method was highly inappropriate
2061 my $class = ref $self;
2062 $self->throw_exception (<<EOE);
2064 No _dbh_last_insert_id() method found in $class.
2065 Since the method of obtaining the autoincrement id of the last insert
2066 operation varies greatly between different databases, this method must be
2067 individually implemented for every storage class.
2071 sub last_insert_id {
2073 $self->_dbh_last_insert_id ($self->_dbh, @_);
2076 =head2 _native_data_type
2080 =item Arguments: $type_name
2084 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2085 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2086 L<::Sybase::ASE|DBIx::Class::Storage::DBI::Sybase::ASE>.
2088 The default implementation returns C<undef>, implement in your Storage driver if
2089 you need this functionality.
2091 Should map types from other databases to the native RDBMS type, for example
2092 C<VARCHAR2> to C<VARCHAR>.
2094 Types with modifiers should map to the underlying data type. For example,
2095 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2097 Composite types should map to the container type, for example
2098 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2102 sub _native_data_type {
2103 #my ($self, $data_type) = @_;
2107 # Check if placeholders are supported at all
2108 sub _placeholders_supported {
2110 my $dbh = $self->_get_dbh;
2112 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2113 # but it is inaccurate more often than not
2115 local $dbh->{PrintError} = 0;
2116 local $dbh->{RaiseError} = 1;
2117 $dbh->do('select ?', {}, 1);
2122 # Check if placeholders bound to non-string types throw exceptions
2124 sub _typeless_placeholders_supported {
2126 my $dbh = $self->_get_dbh;
2129 local $dbh->{PrintError} = 0;
2130 local $dbh->{RaiseError} = 1;
2131 # this specifically tests a bind that is NOT a string
2132 $dbh->do('select 1 where 1 = ?', {}, 1);
2139 Returns the database driver name.
2144 shift->_get_dbh->{Driver}->{Name};
2147 =head2 bind_attribute_by_data_type
2149 Given a datatype from column info, returns a database specific bind
2150 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2151 let the database planner just handle it.
2153 Generally only needed for special case column types, like bytea in postgres.
2157 sub bind_attribute_by_data_type {
2161 =head2 is_datatype_numeric
2163 Given a datatype from column_info, returns a boolean value indicating if
2164 the current RDBMS considers it a numeric value. This controls how
2165 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2166 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2167 be performed instead of the usual C<eq>.
2171 sub is_datatype_numeric {
2172 my ($self, $dt) = @_;
2174 return 0 unless $dt;
2176 return $dt =~ /^ (?:
2177 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2182 =head2 create_ddl_dir
2186 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2190 Creates a SQL file based on the Schema, for each of the specified
2191 database engines in C<\@databases> in the given directory.
2192 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2194 Given a previous version number, this will also create a file containing
2195 the ALTER TABLE statements to transform the previous schema into the
2196 current one. Note that these statements may contain C<DROP TABLE> or
2197 C<DROP COLUMN> statements that can potentially destroy data.
2199 The file names are created using the C<ddl_filename> method below, please
2200 override this method in your schema if you would like a different file
2201 name format. For the ALTER file, the same format is used, replacing
2202 $version in the name with "$preversion-$version".
2204 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2205 The most common value for this would be C<< { add_drop_table => 1 } >>
2206 to have the SQL produced include a C<DROP TABLE> statement for each table
2207 created. For quoting purposes supply C<quote_table_names> and
2208 C<quote_field_names>.
2210 If no arguments are passed, then the following default values are assumed:
2214 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2216 =item version - $schema->schema_version
2218 =item directory - './'
2220 =item preversion - <none>
2224 By default, C<\%sqlt_args> will have
2226 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2228 merged with the hash passed in. To disable any of those features, pass in a
2229 hashref like the following
2231 { ignore_constraint_names => 0, # ... other options }
2234 WARNING: You are strongly advised to check all SQL files created, before applying
2239 sub create_ddl_dir {
2240 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2242 if(!$dir || !-d $dir) {
2243 carp "No directory given, using ./\n";
2246 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2247 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2249 my $schema_version = $schema->schema_version || '1.x';
2250 $version ||= $schema_version;
2253 add_drop_table => 1,
2254 ignore_constraint_names => 1,
2255 ignore_index_names => 1,
2259 $self->throw_exception("Can't create a ddl file without SQL::Translator: " . $self->_sqlt_version_error)
2260 if !$self->_sqlt_version_ok;
2262 my $sqlt = SQL::Translator->new( $sqltargs );
2264 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2265 my $sqlt_schema = $sqlt->translate({ data => $schema })
2266 or $self->throw_exception ($sqlt->error);
2268 foreach my $db (@$databases) {
2270 $sqlt->{schema} = $sqlt_schema;
2271 $sqlt->producer($db);
2274 my $filename = $schema->ddl_filename($db, $version, $dir);
2275 if (-e $filename && ($version eq $schema_version )) {
2276 # if we are dumping the current version, overwrite the DDL
2277 carp "Overwriting existing DDL file - $filename";
2281 my $output = $sqlt->translate;
2283 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2286 if(!open($file, ">$filename")) {
2287 $self->throw_exception("Can't open $filename for writing ($!)");
2290 print $file $output;
2293 next unless ($preversion);
2295 require SQL::Translator::Diff;
2297 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2298 if(!-e $prefilename) {
2299 carp("No previous schema file found ($prefilename)");
2303 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2305 carp("Overwriting existing diff file - $difffile");
2311 my $t = SQL::Translator->new($sqltargs);
2316 or $self->throw_exception ($t->error);
2318 my $out = $t->translate( $prefilename )
2319 or $self->throw_exception ($t->error);
2321 $source_schema = $t->schema;
2323 $source_schema->name( $prefilename )
2324 unless ( $source_schema->name );
2327 # The "new" style of producers have sane normalization and can support
2328 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2329 # And we have to diff parsed SQL against parsed SQL.
2330 my $dest_schema = $sqlt_schema;
2332 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2333 my $t = SQL::Translator->new($sqltargs);
2338 or $self->throw_exception ($t->error);
2340 my $out = $t->translate( $filename )
2341 or $self->throw_exception ($t->error);
2343 $dest_schema = $t->schema;
2345 $dest_schema->name( $filename )
2346 unless $dest_schema->name;
2349 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2353 if(!open $file, ">$difffile") {
2354 $self->throw_exception("Can't write to $difffile ($!)");
2362 =head2 deployment_statements
2366 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2370 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2372 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2373 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2375 C<$directory> is used to return statements from files in a previously created
2376 L</create_ddl_dir> directory and is optional. The filenames are constructed
2377 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2379 If no C<$directory> is specified then the statements are constructed on the
2380 fly using L<SQL::Translator> and C<$version> is ignored.
2382 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2386 sub deployment_statements {
2387 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2388 $type ||= $self->sqlt_type;
2389 $version ||= $schema->schema_version || '1.x';
2391 my $filename = $schema->ddl_filename($type, $version, $dir);
2395 open($file, "<$filename")
2396 or $self->throw_exception("Can't open $filename ($!)");
2399 return join('', @rows);
2402 $self->throw_exception("Can't deploy without either SQL::Translator or a ddl_dir: " . $self->_sqlt_version_error )
2403 if !$self->_sqlt_version_ok;
2405 # sources needs to be a parser arg, but for simplicty allow at top level
2407 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2408 if exists $sqltargs->{sources};
2410 my $tr = SQL::Translator->new(
2411 producer => "SQL::Translator::Producer::${type}",
2413 parser => 'SQL::Translator::Parser::DBIx::Class',
2420 @ret = $tr->translate;
2423 $ret[0] = $tr->translate;
2426 $self->throw_exception( 'Unable to produce deployment statements: ' . $tr->error)
2427 unless (@ret && defined $ret[0]);
2429 return $wa ? @ret : $ret[0];
2433 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2436 return if($line =~ /^--/);
2438 # next if($line =~ /^DROP/m);
2439 return if($line =~ /^BEGIN TRANSACTION/m);
2440 return if($line =~ /^COMMIT/m);
2441 return if $line =~ /^\s+$/; # skip whitespace only
2442 $self->_query_start($line);
2444 # do a dbh_do cycle here, as we need some error checking in
2445 # place (even though we will ignore errors)
2446 $self->dbh_do (sub { $_[1]->do($line) });
2449 carp qq{$@ (running "${line}")};
2451 $self->_query_end($line);
2453 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2454 if (@statements > 1) {
2455 foreach my $statement (@statements) {
2456 $deploy->( $statement );
2459 elsif (@statements == 1) {
2460 foreach my $line ( split(";\n", $statements[0])) {
2466 =head2 datetime_parser
2468 Returns the datetime parser class
2472 sub datetime_parser {
2474 return $self->{datetime_parser} ||= do {
2475 $self->build_datetime_parser(@_);
2479 =head2 datetime_parser_type
2481 Defines (returns) the datetime parser class - currently hardwired to
2482 L<DateTime::Format::MySQL>
2486 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2488 =head2 build_datetime_parser
2490 See L</datetime_parser>
2494 sub build_datetime_parser {
2496 my $type = $self->datetime_parser_type(@_);
2497 $self->ensure_class_loaded ($type);
2502 =head2 is_replicating
2504 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2505 replicate from a master database. Default is undef, which is the result
2506 returned by databases that don't support replication.
2510 sub is_replicating {
2515 =head2 lag_behind_master
2517 Returns a number that represents a certain amount of lag behind a master db
2518 when a given storage is replicating. The number is database dependent, but
2519 starts at zero and increases with the amount of lag. Default in undef
2523 sub lag_behind_master {
2527 # SQLT version handling
2529 my $_sqlt_version_ok; # private
2530 my $_sqlt_version_error; # private
2532 sub _sqlt_version_ok {
2533 if (!defined $_sqlt_version_ok) {
2534 eval "use SQL::Translator $minimum_sqlt_version";
2536 $_sqlt_version_ok = 0;
2537 $_sqlt_version_error = $@;
2540 $_sqlt_version_ok = 1;
2543 return $_sqlt_version_ok;
2546 sub _sqlt_version_error {
2547 shift->_sqlt_version_ok unless defined $_sqlt_version_ok;
2548 return $_sqlt_version_error;
2551 sub _sqlt_minimum_version { $minimum_sqlt_version };
2554 =head2 relname_to_table_alias
2558 =item Arguments: $relname, $join_count
2562 L<DBIx::Class> uses L<DBIx::Class::Relationship> names as table aliases in
2565 This hook is to allow specific L<DBIx::Class::Storage> drivers to change the
2566 way these aliases are named.
2568 The default behavior is C<"$relname_$join_count" if $join_count > 1>, otherwise
2573 sub relname_to_table_alias {
2574 my ($self, $relname, $join_count) = @_;
2576 my $alias = ($join_count && $join_count > 1 ?
2577 join('_', $relname, $join_count) : $relname);
2585 $self->_verify_pid if $self->_dbh;
2587 # some databases need this to stop spewing warnings
2588 if (my $dbh = $self->_dbh) {
2591 %{ $dbh->{CachedKids} } = ();
2603 =head2 DBIx::Class and AutoCommit
2605 DBIx::Class can do some wonderful magic with handling exceptions,
2606 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2607 (the default) combined with C<txn_do> for transaction support.
2609 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2610 in an assumed transaction between commits, and you're telling us you'd
2611 like to manage that manually. A lot of the magic protections offered by
2612 this module will go away. We can't protect you from exceptions due to database
2613 disconnects because we don't know anything about how to restart your
2614 transactions. You're on your own for handling all sorts of exceptional
2615 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2621 Matt S. Trout <mst@shadowcatsystems.co.uk>
2623 Andy Grundman <andy@hybridized.org>
2627 You may distribute this code under the same terms as Perl itself.