1 package DBIx::Class::Storage::DBI;
2 # -*- mode: cperl; cperl-indent-level: 2 -*-
7 use base '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();
18 # what version of sqlt do we require if deploy() without a ddl_dir is invoked
19 # when changing also adjust the corresponding author_require in Makefile.PL
20 my $minimum_sqlt_version = '0.11002';
23 __PACKAGE__->mk_group_accessors('simple' =>
24 qw/_connect_info _dbi_connect_info _dbh _sql_maker _sql_maker_opts _conn_pid
25 _conn_tid transaction_depth _dbh_autocommit _driver_determined savepoints/
28 # the values for these accessors are picked out (and deleted) from
29 # the attribute hashref passed to connect_info
30 my @storage_options = qw/
31 on_connect_call on_disconnect_call on_connect_do on_disconnect_do
32 disable_sth_caching unsafe auto_savepoint
34 __PACKAGE__->mk_group_accessors('simple' => @storage_options);
37 # default cursor class, overridable in connect_info attributes
38 __PACKAGE__->cursor_class('DBIx::Class::Storage::DBI::Cursor');
40 __PACKAGE__->mk_group_accessors('inherited' => qw/sql_maker_class/);
41 __PACKAGE__->sql_maker_class('DBIx::Class::SQLAHacks');
46 DBIx::Class::Storage::DBI - DBI storage handler
50 my $schema = MySchema->connect('dbi:SQLite:my.db');
52 $schema->storage->debug(1);
54 my @stuff = $schema->storage->dbh_do(
56 my ($storage, $dbh, @args) = @_;
57 $dbh->do("DROP TABLE authors");
62 $schema->resultset('Book')->search({
63 written_on => $schema->storage->datetime_parser(DateTime->now)
68 This class represents the connection to an RDBMS via L<DBI>. See
69 L<DBIx::Class::Storage> for general information. This pod only
70 documents DBI-specific methods and behaviors.
77 my $new = shift->next::method(@_);
79 $new->transaction_depth(0);
80 $new->_sql_maker_opts({});
81 $new->{savepoints} = [];
82 $new->{_in_dbh_do} = 0;
90 This method is normally called by L<DBIx::Class::Schema/connection>, which
91 encapsulates its argument list in an arrayref before passing them here.
93 The argument list may contain:
99 The same 4-element argument set one would normally pass to
100 L<DBI/connect>, optionally followed by
101 L<extra attributes|/DBIx::Class specific connection attributes>
102 recognized by DBIx::Class:
104 $connect_info_args = [ $dsn, $user, $password, \%dbi_attributes?, \%extra_attributes? ];
108 A single code reference which returns a connected
109 L<DBI database handle|DBI/connect> optionally followed by
110 L<extra attributes|/DBIx::Class specific connection attributes> recognized
113 $connect_info_args = [ sub { DBI->connect (...) }, \%extra_attributes? ];
117 A single hashref with all the attributes and the dsn/user/password
120 $connect_info_args = [{
128 $connect_info_args = [{
129 dbh_maker => sub { DBI->connect (...) },
134 This is particularly useful for L<Catalyst> based applications, allowing the
135 following config (L<Config::General> style):
140 dsn dbi:mysql:database=test
147 The C<dsn>/C<user>/C<password> combination can be substituted by the
148 C<dbh_maker> key whose value is a coderef that returns a connected
149 L<DBI database handle|DBI/connect>
153 Please note that the L<DBI> docs recommend that you always explicitly
154 set C<AutoCommit> to either I<0> or I<1>. L<DBIx::Class> further
155 recommends that it be set to I<1>, and that you perform transactions
156 via our L<DBIx::Class::Schema/txn_do> method. L<DBIx::Class> will set it
157 to I<1> if you do not do explicitly set it to zero. This is the default
158 for most DBDs. See L</DBIx::Class and AutoCommit> for details.
160 =head3 DBIx::Class specific connection attributes
162 In addition to the standard L<DBI|DBI/ATTRIBUTES_COMMON_TO_ALL_HANDLES>
163 L<connection|DBI/Database_Handle_Attributes> attributes, DBIx::Class recognizes
164 the following connection options. These options can be mixed in with your other
165 L<DBI> connection attributes, or placed in a seperate hashref
166 (C<\%extra_attributes>) as shown above.
168 Every time C<connect_info> is invoked, any previous settings for
169 these options will be cleared before setting the new ones, regardless of
170 whether any options are specified in the new C<connect_info>.
177 Specifies things to do immediately after connecting or re-connecting to
178 the database. Its value may contain:
184 This contains one SQL statement to execute.
186 =item an array reference
188 This contains SQL statements to execute in order. Each element contains
189 a string or a code reference that returns a string.
191 =item a code reference
193 This contains some code to execute. Unlike code references within an
194 array reference, its return value is ignored.
198 =item on_disconnect_do
200 Takes arguments in the same form as L</on_connect_do> and executes them
201 immediately before disconnecting from the database.
203 Note, this only runs if you explicitly call L</disconnect> on the
206 =item on_connect_call
208 A more generalized form of L</on_connect_do> that calls the specified
209 C<connect_call_METHOD> methods in your storage driver.
211 on_connect_do => 'select 1'
215 on_connect_call => [ [ do_sql => 'select 1' ] ]
217 Its values may contain:
223 Will call the C<connect_call_METHOD> method.
225 =item a code reference
227 Will execute C<< $code->($storage) >>
229 =item an array reference
231 Each value can be a method name or code reference.
233 =item an array of arrays
235 For each array, the first item is taken to be the C<connect_call_> method name
236 or code reference, and the rest are parameters to it.
240 Some predefined storage methods you may use:
246 Executes a SQL string or a code reference that returns a SQL string. This is
247 what L</on_connect_do> and L</on_disconnect_do> use.
255 Will execute the scalar as SQL.
259 Taken to be arguments to L<DBI/do>, the SQL string optionally followed by the
260 attributes hashref and bind values.
262 =item a code reference
264 Will execute C<< $code->($storage) >> and execute the return array refs as
271 Execute any statements necessary to initialize the database session to return
272 and accept datetime/timestamp values used with
273 L<DBIx::Class::InflateColumn::DateTime>.
275 Only necessary for some databases, see your specific storage driver for
276 implementation details.
280 =item on_disconnect_call
282 Takes arguments in the same form as L</on_connect_call> and executes them
283 immediately before disconnecting from the database.
285 Calls the C<disconnect_call_METHOD> methods as opposed to the
286 C<connect_call_METHOD> methods called by L</on_connect_call>.
288 Note, this only runs if you explicitly call L</disconnect> on the
291 =item disable_sth_caching
293 If set to a true value, this option will disable the caching of
294 statement handles via L<DBI/prepare_cached>.
298 Sets the limit dialect. This is useful for JDBC-bridge among others
299 where the remote SQL-dialect cannot be determined by the name of the
300 driver alone. See also L<SQL::Abstract::Limit>.
304 Specifies what characters to use to quote table and column names. If
305 you use this you will want to specify L</name_sep> as well.
307 C<quote_char> expects either a single character, in which case is it
308 is placed on either side of the table/column name, or an arrayref of length
309 2 in which case the table/column name is placed between the elements.
311 For example under MySQL you should use C<< quote_char => '`' >>, and for
312 SQL Server you should use C<< quote_char => [qw/[ ]/] >>.
316 This only needs to be used in conjunction with C<quote_char>, and is used to
317 specify the charecter that seperates elements (schemas, tables, columns) from
318 each other. In most cases this is simply a C<.>.
320 The consequences of not supplying this value is that L<SQL::Abstract>
321 will assume DBIx::Class' uses of aliases to be complete column
322 names. The output will look like I<"me.name"> when it should actually
327 This Storage driver normally installs its own C<HandleError>, sets
328 C<RaiseError> and C<ShowErrorStatement> on, and sets C<PrintError> off on
329 all database handles, including those supplied by a coderef. It does this
330 so that it can have consistent and useful error behavior.
332 If you set this option to a true value, Storage will not do its usual
333 modifications to the database handle's attributes, and instead relies on
334 the settings in your connect_info DBI options (or the values you set in
335 your connection coderef, in the case that you are connecting via coderef).
337 Note that your custom settings can cause Storage to malfunction,
338 especially if you set a C<HandleError> handler that suppresses exceptions
339 and/or disable C<RaiseError>.
343 If this option is true, L<DBIx::Class> will use savepoints when nesting
344 transactions, making it possible to recover from failure in the inner
345 transaction without having to abort all outer transactions.
349 Use this argument to supply a cursor class other than the default
350 L<DBIx::Class::Storage::DBI::Cursor>.
354 Some real-life examples of arguments to L</connect_info> and
355 L<DBIx::Class::Schema/connect>
357 # Simple SQLite connection
358 ->connect_info([ 'dbi:SQLite:./foo.db' ]);
361 ->connect_info([ sub { DBI->connect(...) } ]);
363 # Connect via subref in hashref
365 dbh_maker => sub { DBI->connect(...) },
366 on_connect_do => 'alter session ...',
369 # A bit more complicated
376 { quote_char => q{"}, name_sep => q{.} },
380 # Equivalent to the previous example
386 { AutoCommit => 1, quote_char => q{"}, name_sep => q{.} },
390 # Same, but with hashref as argument
391 # See parse_connect_info for explanation
394 dsn => 'dbi:Pg:dbname=foo',
396 password => 'my_pg_password',
403 # Subref + DBIx::Class-specific connection options
406 sub { DBI->connect(...) },
410 on_connect_do => ['SET search_path TO myschema,otherschema,public'],
411 disable_sth_caching => 1,
421 my ($self, $info_arg) = @_;
423 return $self->_connect_info if !$info_arg;
425 my @args = @$info_arg; # take a shallow copy for further mutilation
426 $self->_connect_info([@args]); # copy for _connect_info
429 # combine/pre-parse arguments depending on invocation style
432 if (ref $args[0] eq 'CODE') { # coderef with optional \%extra_attributes
433 %attrs = %{ $args[1] || {} };
436 elsif (ref $args[0] eq 'HASH') { # single hashref (i.e. Catalyst config)
437 %attrs = %{$args[0]};
439 if (my $code = delete $attrs{dbh_maker}) {
442 my @ignored = grep { delete $attrs{$_} } (qw/dsn user password/);
445 'Attribute(s) %s in connect_info were ignored, as they can not be applied '
446 . "to the result of 'dbh_maker'",
448 join (', ', map { "'$_'" } (@ignored) ),
453 @args = delete @attrs{qw/dsn user password/};
456 else { # otherwise assume dsn/user/password + \%attrs + \%extra_attrs
458 % { $args[3] || {} },
459 % { $args[4] || {} },
461 @args = @args[0,1,2];
464 # Kill sql_maker/_sql_maker_opts, so we get a fresh one with only
465 # the new set of options
466 $self->_sql_maker(undef);
467 $self->_sql_maker_opts({});
470 for my $storage_opt (@storage_options, 'cursor_class') { # @storage_options is declared at the top of the module
471 if(my $value = delete $attrs{$storage_opt}) {
472 $self->$storage_opt($value);
475 for my $sql_maker_opt (qw/limit_dialect quote_char name_sep/) {
476 if(my $opt_val = delete $attrs{$sql_maker_opt}) {
477 $self->_sql_maker_opts->{$sql_maker_opt} = $opt_val;
482 if (ref $args[0] eq 'CODE') {
483 # _connect() never looks past $args[0] in this case
487 %{ $self->_default_dbi_connect_attributes || {} },
492 $self->_dbi_connect_info([@args, keys %attrs ? \%attrs : ()]);
493 $self->_connect_info;
496 sub _default_dbi_connect_attributes {
506 This method is deprecated in favour of setting via L</connect_info>.
510 =head2 on_disconnect_do
512 This method is deprecated in favour of setting via L</connect_info>.
516 sub _parse_connect_do {
517 my ($self, $type) = @_;
519 my $val = $self->$type;
520 return () if not defined $val;
525 push @res, [ 'do_sql', $val ];
526 } elsif (ref($val) eq 'CODE') {
528 } elsif (ref($val) eq 'ARRAY') {
529 push @res, map { [ 'do_sql', $_ ] } @$val;
531 $self->throw_exception("Invalid type for $type: ".ref($val));
539 Arguments: ($subref | $method_name), @extra_coderef_args?
541 Execute the given $subref or $method_name using the new exception-based
542 connection management.
544 The first two arguments will be the storage object that C<dbh_do> was called
545 on and a database handle to use. Any additional arguments will be passed
546 verbatim to the called subref as arguments 2 and onwards.
548 Using this (instead of $self->_dbh or $self->dbh) ensures correct
549 exception handling and reconnection (or failover in future subclasses).
551 Your subref should have no side-effects outside of the database, as
552 there is the potential for your subref to be partially double-executed
553 if the database connection was stale/dysfunctional.
557 my @stuff = $schema->storage->dbh_do(
559 my ($storage, $dbh, @cols) = @_;
560 my $cols = join(q{, }, @cols);
561 $dbh->selectrow_array("SELECT $cols FROM foo");
572 my $dbh = $self->_get_dbh;
574 return $self->$code($dbh, @_) if $self->{_in_dbh_do}
575 || $self->{transaction_depth};
577 local $self->{_in_dbh_do} = 1;
580 my $want_array = wantarray;
585 @result = $self->$code($dbh, @_);
587 elsif(defined $want_array) {
588 $result[0] = $self->$code($dbh, @_);
591 $self->$code($dbh, @_);
595 # ->connected might unset $@ - copy
597 if(!$exception) { return $want_array ? @result : $result[0] }
599 $self->throw_exception($exception) if $self->connected;
601 # We were not connected - reconnect and retry, but let any
602 # exception fall right through this time
603 carp "Retrying $code after catching disconnected exception: $exception"
604 if $ENV{DBIC_DBIRETRY_DEBUG};
605 $self->_populate_dbh;
606 $self->$code($self->_dbh, @_);
609 # This is basically a blend of dbh_do above and DBIx::Class::Storage::txn_do.
610 # It also informs dbh_do to bypass itself while under the direction of txn_do,
611 # via $self->{_in_dbh_do} (this saves some redundant eval and errorcheck, etc)
616 ref $coderef eq 'CODE' or $self->throw_exception
617 ('$coderef must be a CODE reference');
619 return $coderef->(@_) if $self->{transaction_depth} && ! $self->auto_savepoint;
621 local $self->{_in_dbh_do} = 1;
624 my $want_array = wantarray;
633 @result = $coderef->(@_);
635 elsif(defined $want_array) {
636 $result[0] = $coderef->(@_);
644 # ->connected might unset $@ - copy
646 if(!$exception) { return $want_array ? @result : $result[0] }
648 if($tried++ || $self->connected) {
649 eval { $self->txn_rollback };
650 my $rollback_exception = $@;
651 if($rollback_exception) {
652 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
653 $self->throw_exception($exception) # propagate nested rollback
654 if $rollback_exception =~ /$exception_class/;
656 $self->throw_exception(
657 "Transaction aborted: ${exception}. "
658 . "Rollback failed: ${rollback_exception}"
661 $self->throw_exception($exception)
664 # We were not connected, and was first try - reconnect and retry
666 carp "Retrying $coderef after catching disconnected exception: $exception"
667 if $ENV{DBIC_DBIRETRY_DEBUG};
668 $self->_populate_dbh;
674 Our C<disconnect> method also performs a rollback first if the
675 database is not in C<AutoCommit> mode.
685 push @actions, ( $self->on_disconnect_call || () );
686 push @actions, $self->_parse_connect_do ('on_disconnect_do');
688 $self->_do_connection_actions(disconnect_call_ => $_) for @actions;
690 $self->_dbh_rollback unless $self->_dbh_autocommit;
692 $self->_dbh->disconnect;
698 =head2 with_deferred_fk_checks
702 =item Arguments: C<$coderef>
704 =item Return Value: The return value of $coderef
708 Storage specific method to run the code ref with FK checks deferred or
709 in MySQL's case disabled entirely.
713 # Storage subclasses should override this
714 sub with_deferred_fk_checks {
715 my ($self, $sub) = @_;
724 =item Arguments: none
726 =item Return Value: 1|0
730 Verifies that the the current database handle is active and ready to execute
731 an SQL statement (i.e. the connection did not get stale, server is still
732 answering, etc.) This method is used internally by L</dbh>.
738 return 0 unless $self->_seems_connected;
741 local $self->_dbh->{RaiseError} = 1;
746 sub _seems_connected {
749 my $dbh = $self->_dbh
752 if(defined $self->_conn_tid && $self->_conn_tid != threads->tid) {
759 return 0 if !$self->_dbh;
762 return $dbh->FETCH('Active');
768 my $dbh = $self->_dbh or return 0;
773 # handle pid changes correctly
774 # NOTE: assumes $self->_dbh is a valid $dbh
778 return if defined $self->_conn_pid && $self->_conn_pid == $$;
780 $self->_dbh->{InactiveDestroy} = 1;
787 sub ensure_connected {
790 unless ($self->connected) {
791 $self->_populate_dbh;
797 Returns a C<$dbh> - a data base handle of class L<DBI>. The returned handle
798 is guaranteed to be healthy by implicitly calling L</connected>, and if
799 necessary performing a reconnection before returning. Keep in mind that this
800 is very B<expensive> on some database engines. Consider using L<dbh_do>
808 if (not $self->_dbh) {
809 $self->_populate_dbh;
811 $self->ensure_connected;
816 # this is the internal "get dbh or connect (don't check)" method
819 $self->_verify_pid if $self->_dbh;
820 $self->_populate_dbh unless $self->_dbh;
824 sub _sql_maker_args {
829 array_datatypes => 1,
830 limit_dialect => $self->_get_dbh,
831 %{$self->_sql_maker_opts}
837 unless ($self->_sql_maker) {
838 my $sql_maker_class = $self->sql_maker_class;
839 $self->ensure_class_loaded ($sql_maker_class);
840 $self->_sql_maker($sql_maker_class->new( $self->_sql_maker_args ));
842 return $self->_sql_maker;
845 # nothing to do by default
852 my @info = @{$self->_dbi_connect_info || []};
853 $self->_dbh(undef); # in case ->connected failed we might get sent here
854 $self->_dbh($self->_connect(@info));
856 $self->_conn_pid($$);
857 $self->_conn_tid(threads->tid) if $INC{'threads.pm'};
859 $self->_determine_driver;
861 # Always set the transaction depth on connect, since
862 # there is no transaction in progress by definition
863 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
865 $self->_run_connection_actions unless $self->{_in_determine_driver};
868 sub _run_connection_actions {
872 push @actions, ( $self->on_connect_call || () );
873 push @actions, $self->_parse_connect_do ('on_connect_do');
875 $self->_do_connection_actions(connect_call_ => $_) for @actions;
878 sub _determine_driver {
881 if ((not $self->_driver_determined) && (not $self->{_in_determine_driver})) {
882 my $started_connected = 0;
883 local $self->{_in_determine_driver} = 1;
885 if (ref($self) eq __PACKAGE__) {
887 if ($self->_dbh) { # we are connected
888 $driver = $self->_dbh->{Driver}{Name};
889 $started_connected = 1;
891 # if connect_info is a CODEREF, we have no choice but to connect
892 if (ref $self->_dbi_connect_info->[0] &&
893 Scalar::Util::reftype($self->_dbi_connect_info->[0]) eq 'CODE') {
894 $self->_populate_dbh;
895 $driver = $self->_dbh->{Driver}{Name};
898 # try to use dsn to not require being connected, the driver may still
899 # force a connection in _rebless to determine version
900 ($driver) = $self->_dbi_connect_info->[0] =~ /dbi:([^:]+):/i;
904 my $storage_class = "DBIx::Class::Storage::DBI::${driver}";
905 if ($self->load_optional_class($storage_class)) {
906 mro::set_mro($storage_class, 'c3');
907 bless $self, $storage_class;
912 $self->_driver_determined(1);
914 $self->_init; # run driver-specific initializations
916 $self->_run_connection_actions
917 if !$started_connected && defined $self->_dbh;
921 sub _do_connection_actions {
923 my $method_prefix = shift;
926 if (not ref($call)) {
927 my $method = $method_prefix . $call;
929 } elsif (ref($call) eq 'CODE') {
931 } elsif (ref($call) eq 'ARRAY') {
932 if (ref($call->[0]) ne 'ARRAY') {
933 $self->_do_connection_actions($method_prefix, $_) for @$call;
935 $self->_do_connection_actions($method_prefix, @$_) for @$call;
938 $self->throw_exception (sprintf ("Don't know how to process conection actions of type '%s'", ref($call)) );
944 sub connect_call_do_sql {
946 $self->_do_query(@_);
949 sub disconnect_call_do_sql {
951 $self->_do_query(@_);
954 # override in db-specific backend when necessary
955 sub connect_call_datetime_setup { 1 }
958 my ($self, $action) = @_;
960 if (ref $action eq 'CODE') {
961 $action = $action->($self);
962 $self->_do_query($_) foreach @$action;
965 # Most debuggers expect ($sql, @bind), so we need to exclude
966 # the attribute hash which is the second argument to $dbh->do
967 # furthermore the bind values are usually to be presented
968 # as named arrayref pairs, so wrap those here too
969 my @do_args = (ref $action eq 'ARRAY') ? (@$action) : ($action);
970 my $sql = shift @do_args;
971 my $attrs = shift @do_args;
972 my @bind = map { [ undef, $_ ] } @do_args;
974 $self->_query_start($sql, @bind);
975 $self->_get_dbh->do($sql, $attrs, @do_args);
976 $self->_query_end($sql, @bind);
983 my ($self, @info) = @_;
985 $self->throw_exception("You failed to provide any connection info")
988 my ($old_connect_via, $dbh);
990 if ($INC{'Apache/DBI.pm'} && $ENV{MOD_PERL}) {
991 $old_connect_via = $DBI::connect_via;
992 $DBI::connect_via = 'connect';
996 if(ref $info[0] eq 'CODE') {
1000 $dbh = DBI->connect(@info);
1003 if($dbh && !$self->unsafe) {
1004 my $weak_self = $self;
1005 Scalar::Util::weaken($weak_self);
1006 $dbh->{HandleError} = sub {
1008 $weak_self->throw_exception("DBI Exception: $_[0]");
1011 # the handler may be invoked by something totally out of
1013 croak ("DBI Exception: $_[0]");
1016 $dbh->{ShowErrorStatement} = 1;
1017 $dbh->{RaiseError} = 1;
1018 $dbh->{PrintError} = 0;
1022 $DBI::connect_via = $old_connect_via if $old_connect_via;
1024 $self->throw_exception("DBI Connection failed: " . ($@||$DBI::errstr))
1027 $self->_dbh_autocommit($dbh->{AutoCommit});
1033 my ($self, $name) = @_;
1035 $name = $self->_svp_generate_name
1036 unless defined $name;
1038 $self->throw_exception ("You can't use savepoints outside a transaction")
1039 if $self->{transaction_depth} == 0;
1041 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1042 unless $self->can('_svp_begin');
1044 push @{ $self->{savepoints} }, $name;
1046 $self->debugobj->svp_begin($name) if $self->debug;
1048 return $self->_svp_begin($name);
1052 my ($self, $name) = @_;
1054 $self->throw_exception ("You can't use savepoints outside a transaction")
1055 if $self->{transaction_depth} == 0;
1057 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1058 unless $self->can('_svp_release');
1060 if (defined $name) {
1061 $self->throw_exception ("Savepoint '$name' does not exist")
1062 unless grep { $_ eq $name } @{ $self->{savepoints} };
1064 # Dig through the stack until we find the one we are releasing. This keeps
1065 # the stack up to date.
1068 do { $svp = pop @{ $self->{savepoints} } } while $svp ne $name;
1070 $name = pop @{ $self->{savepoints} };
1073 $self->debugobj->svp_release($name) if $self->debug;
1075 return $self->_svp_release($name);
1079 my ($self, $name) = @_;
1081 $self->throw_exception ("You can't use savepoints outside a transaction")
1082 if $self->{transaction_depth} == 0;
1084 $self->throw_exception ("Your Storage implementation doesn't support savepoints")
1085 unless $self->can('_svp_rollback');
1087 if (defined $name) {
1088 # If they passed us a name, verify that it exists in the stack
1089 unless(grep({ $_ eq $name } @{ $self->{savepoints} })) {
1090 $self->throw_exception("Savepoint '$name' does not exist!");
1093 # Dig through the stack until we find the one we are releasing. This keeps
1094 # the stack up to date.
1095 while(my $s = pop(@{ $self->{savepoints} })) {
1096 last if($s eq $name);
1098 # Add the savepoint back to the stack, as a rollback doesn't remove the
1099 # named savepoint, only everything after it.
1100 push(@{ $self->{savepoints} }, $name);
1102 # We'll assume they want to rollback to the last savepoint
1103 $name = $self->{savepoints}->[-1];
1106 $self->debugobj->svp_rollback($name) if $self->debug;
1108 return $self->_svp_rollback($name);
1111 sub _svp_generate_name {
1114 return 'savepoint_'.scalar(@{ $self->{'savepoints'} });
1119 if($self->{transaction_depth} == 0) {
1120 $self->debugobj->txn_begin()
1122 $self->_dbh_begin_work;
1124 elsif ($self->auto_savepoint) {
1127 $self->{transaction_depth}++;
1130 sub _dbh_begin_work {
1133 # if the user is utilizing txn_do - good for him, otherwise we need to
1134 # ensure that the $dbh is healthy on BEGIN.
1135 # We do this via ->dbh_do instead of ->dbh, so that the ->dbh "ping"
1136 # will be replaced by a failure of begin_work itself (which will be
1137 # then retried on reconnect)
1138 if ($self->{_in_dbh_do}) {
1139 $self->_dbh->begin_work;
1141 $self->dbh_do(sub { $_[1]->begin_work });
1147 if ($self->{transaction_depth} == 1) {
1148 my $dbh = $self->_dbh;
1149 $self->debugobj->txn_commit()
1152 $self->{transaction_depth} = 0
1153 if $self->_dbh_autocommit;
1155 elsif($self->{transaction_depth} > 1) {
1156 $self->{transaction_depth}--;
1158 if $self->auto_savepoint;
1164 $self->_dbh->commit;
1169 my $dbh = $self->_dbh;
1171 if ($self->{transaction_depth} == 1) {
1172 $self->debugobj->txn_rollback()
1174 $self->{transaction_depth} = 0
1175 if $self->_dbh_autocommit;
1176 $self->_dbh_rollback;
1178 elsif($self->{transaction_depth} > 1) {
1179 $self->{transaction_depth}--;
1180 if ($self->auto_savepoint) {
1181 $self->svp_rollback;
1186 die DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION->new;
1191 my $exception_class = "DBIx::Class::Storage::NESTED_ROLLBACK_EXCEPTION";
1192 $error =~ /$exception_class/ and $self->throw_exception($error);
1193 # ensure that a failed rollback resets the transaction depth
1194 $self->{transaction_depth} = $self->_dbh_autocommit ? 0 : 1;
1195 $self->throw_exception($error);
1201 $self->_dbh->rollback;
1204 # This used to be the top-half of _execute. It was split out to make it
1205 # easier to override in NoBindVars without duping the rest. It takes up
1206 # all of _execute's args, and emits $sql, @bind.
1207 sub _prep_for_execute {
1208 my ($self, $op, $extra_bind, $ident, $args) = @_;
1210 if( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
1211 $ident = $ident->from();
1214 my ($sql, @bind) = $self->sql_maker->$op($ident, @$args);
1217 map { ref $_ eq 'ARRAY' ? $_ : [ '!!dummy', $_ ] } @$extra_bind)
1219 return ($sql, \@bind);
1223 sub _fix_bind_params {
1224 my ($self, @bind) = @_;
1226 ### Turn @bind from something like this:
1227 ### ( [ "artist", 1 ], [ "cdid", 1, 3 ] )
1229 ### ( "'1'", "'1'", "'3'" )
1232 if ( defined( $_ && $_->[1] ) ) {
1233 map { qq{'$_'}; } @{$_}[ 1 .. $#$_ ];
1240 my ( $self, $sql, @bind ) = @_;
1242 if ( $self->debug ) {
1243 @bind = $self->_fix_bind_params(@bind);
1245 $self->debugobj->query_start( $sql, @bind );
1250 my ( $self, $sql, @bind ) = @_;
1252 if ( $self->debug ) {
1253 @bind = $self->_fix_bind_params(@bind);
1254 $self->debugobj->query_end( $sql, @bind );
1259 my ($self, $dbh, $op, $extra_bind, $ident, $bind_attributes, @args) = @_;
1261 my ($sql, $bind) = $self->_prep_for_execute($op, $extra_bind, $ident, \@args);
1263 $self->_query_start( $sql, @$bind );
1265 my $sth = $self->sth($sql,$op);
1267 my $placeholder_index = 1;
1269 foreach my $bound (@$bind) {
1270 my $attributes = {};
1271 my($column_name, @data) = @$bound;
1273 if ($bind_attributes) {
1274 $attributes = $bind_attributes->{$column_name}
1275 if defined $bind_attributes->{$column_name};
1278 foreach my $data (@data) {
1279 my $ref = ref $data;
1280 $data = $ref && $ref ne 'ARRAY' ? ''.$data : $data; # stringify args (except arrayrefs)
1282 $sth->bind_param($placeholder_index, $data, $attributes);
1283 $placeholder_index++;
1287 # Can this fail without throwing an exception anyways???
1288 my $rv = $sth->execute();
1289 $self->throw_exception($sth->errstr) if !$rv;
1291 $self->_query_end( $sql, @$bind );
1293 return (wantarray ? ($rv, $sth, @$bind) : $rv);
1298 $self->dbh_do('_dbh_execute', @_); # retry over disconnects
1302 my ($self, $source, $to_insert) = @_;
1304 # redispatch to insert method of storage we reblessed into, if necessary
1305 if (not $self->_driver_determined) {
1306 $self->_determine_driver;
1307 goto $self->can('insert');
1310 my $ident = $source->from;
1311 my $bind_attributes = $self->source_bind_attributes($source);
1313 my $updated_cols = {};
1315 foreach my $col ( $source->columns ) {
1316 if ( !defined $to_insert->{$col} ) {
1317 my $col_info = $source->column_info($col);
1319 if ( $col_info->{auto_nextval} ) {
1320 $updated_cols->{$col} = $to_insert->{$col} = $self->_sequence_fetch(
1322 $col_info->{sequence} ||
1323 $self->_dbh_get_autoinc_seq($self->_get_dbh, $source)
1329 $self->_execute('insert' => [], $source, $bind_attributes, $to_insert);
1331 return $updated_cols;
1334 ## Still not quite perfect, and EXPERIMENTAL
1335 ## Currently it is assumed that all values passed will be "normal", i.e. not
1336 ## scalar refs, or at least, all the same type as the first set, the statement is
1337 ## only prepped once.
1339 my ($self, $source, $cols, $data) = @_;
1341 # redispatch to insert_bulk method of storage we reblessed into, if necessary
1342 if (not $self->_driver_determined) {
1343 $self->_determine_driver;
1344 goto $self->can('insert_bulk');
1348 @colvalues{@$cols} = (0..$#$cols);
1350 for my $i (0..$#$cols) {
1351 my $first_val = $data->[0][$i];
1352 next unless ref $first_val eq 'SCALAR';
1354 $colvalues{ $cols->[$i] } = $first_val;
1355 ## This is probably unnecessary since $rs->populate only looks at the first
1358 # ref $_ eq 'SCALAR' && $$_ eq $$first_val
1359 # } map $data->[$_][$i], (1..$#$data)) == (@$data - 1);
1362 # check for bad data
1363 my $bad_slice = sub {
1364 my ($msg, $col_idx, $slice_idx) = @_;
1365 $self->throw_exception(sprintf "%s for column '%s' in populate slice:\n%s",
1368 Data::Dumper::Concise::Dumper({
1369 map { $cols->[$_] => $data->[$slice_idx][$_] } (0 .. $#$cols)
1374 for my $datum_idx (0..$#$data) {
1375 my $datum = $data->[$datum_idx];
1377 for my $col_idx (0..$#$cols) {
1378 my $val = $datum->[$col_idx];
1379 my $sqla_bind = $colvalues{ $cols->[$col_idx] };
1380 my $is_literal_sql = (ref $sqla_bind) eq 'SCALAR';
1382 if ($is_literal_sql) {
1384 $bad_slice->('bind found where literal SQL expected', $col_idx, $datum_idx);
1386 elsif ((my $reftype = ref $val) ne 'SCALAR') {
1387 $bad_slice->("$reftype reference found where literal SQL expected",
1388 $col_idx, $datum_idx);
1390 elsif ($$val ne $$sqla_bind){
1391 $bad_slice->("inconsistent literal SQL value, expecting: '$$sqla_bind'",
1392 $col_idx, $datum_idx);
1395 elsif (my $reftype = ref $val) {
1396 $bad_slice->("$reftype reference found where bind expected",
1397 $col_idx, $datum_idx);
1402 my ($sql, $bind) = $self->_prep_for_execute (
1403 'insert', undef, $source, [\%colvalues]
1407 my $empty_bind = 1 if (not @bind) &&
1408 (grep { ref $_ eq 'SCALAR' } values %colvalues) == @$cols;
1410 if ((not @bind) && (not $empty_bind)) {
1411 $self->throw_exception(
1412 'Cannot insert_bulk without support for placeholders'
1416 $self->_query_start( $sql, ['__BULK__'] );
1417 my $sth = $self->sth($sql);
1421 # bind_param_array doesn't work if there are no binds
1422 $self->_dbh_execute_inserts_with_no_binds( $sth, scalar @$data );
1425 # @bind = map { ref $_ ? ''.$_ : $_ } @bind; # stringify args
1426 $self->_execute_array( $source, $sth, \@bind, $cols, $data );
1430 $self->_query_end( $sql, ['__BULK__'] );
1432 return (wantarray ? ($rv, $sth, @bind) : $rv);
1435 sub _execute_array {
1436 my ($self, $source, $sth, $bind, $cols, $data, @extra) = @_;
1438 my $guard = $self->txn_scope_guard unless $self->{transaction_depth} != 0;
1440 ## This must be an arrayref, else nothing works!
1441 my $tuple_status = [];
1443 ## Get the bind_attributes, if any exist
1444 my $bind_attributes = $self->source_bind_attributes($source);
1446 ## Bind the values and execute
1447 my $placeholder_index = 1;
1449 foreach my $bound (@$bind) {
1451 my $attributes = {};
1452 my ($column_name, $data_index) = @$bound;
1454 if( $bind_attributes ) {
1455 $attributes = $bind_attributes->{$column_name}
1456 if defined $bind_attributes->{$column_name};
1459 my @data = map { $_->[$data_index] } @$data;
1461 $sth->bind_param_array( $placeholder_index, [@data], $attributes );
1462 $placeholder_index++;
1466 $self->_dbh_execute_array($sth, $tuple_status, @extra);
1468 my $err = $@ || $sth->errstr;
1470 # Statement must finish even if there was an exception.
1471 eval { $sth->finish };
1472 $err = $@ unless $err;
1476 ++$i while $i <= $#$tuple_status && !ref $tuple_status->[$i];
1478 $self->throw_exception("Unexpected populate error: $err")
1479 if ($i > $#$tuple_status);
1481 $self->throw_exception(sprintf "%s for populate slice:\n%s",
1482 ($tuple_status->[$i][1] || $err),
1483 Data::Dumper::Concise::Dumper({
1484 map { $cols->[$_] => $data->[$i][$_] } (0 .. $#$cols)
1489 $guard->commit if $guard;
1494 sub _dbh_execute_array {
1495 my ($self, $sth, $tuple_status, @extra) = @_;
1497 return $sth->execute_array({ArrayTupleStatus => $tuple_status});
1500 sub _dbh_execute_inserts_with_no_binds {
1501 my ($self, $sth, $count) = @_;
1503 my $guard = $self->txn_scope_guard unless $self->{transaction_depth} != 0;
1506 my $dbh = $self->_get_dbh;
1507 local $dbh->{RaiseError} = 1;
1508 local $dbh->{PrintError} = 0;
1510 $sth->execute foreach 1..$count;
1514 # Make sure statement is finished even if there was an exception.
1515 eval { $sth->finish };
1516 $exception = $@ unless $exception;
1518 $self->throw_exception($exception) if $exception;
1520 $guard->commit if $guard;
1526 my ($self, $source, @args) = @_;
1528 # redispatch to update method of storage we reblessed into, if necessary
1529 if (not $self->_driver_determined) {
1530 $self->_determine_driver;
1531 goto $self->can('update');
1534 my $bind_attributes = $self->source_bind_attributes($source);
1536 return $self->_execute('update' => [], $source, $bind_attributes, @args);
1541 my $self = shift @_;
1542 my $source = shift @_;
1543 $self->_determine_driver;
1544 my $bind_attrs = $self->source_bind_attributes($source);
1546 return $self->_execute('delete' => [], $source, $bind_attrs, @_);
1549 # We were sent here because the $rs contains a complex search
1550 # which will require a subquery to select the correct rows
1551 # (i.e. joined or limited resultsets)
1553 # Genarating a single PK column subquery is trivial and supported
1554 # by all RDBMS. However if we have a multicolumn PK, things get ugly.
1555 # Look at _multipk_update_delete()
1556 sub _subq_update_delete {
1558 my ($rs, $op, $values) = @_;
1560 my $rsrc = $rs->result_source;
1562 # we already check this, but double check naively just in case. Should be removed soon
1563 my $sel = $rs->_resolved_attrs->{select};
1564 $sel = [ $sel ] unless ref $sel eq 'ARRAY';
1565 my @pcols = $rsrc->primary_columns;
1566 if (@$sel != @pcols) {
1567 $self->throw_exception (
1568 'Subquery update/delete can not be called on resultsets selecting a'
1569 .' number of columns different than the number of primary keys'
1576 $op eq 'update' ? $values : (),
1577 { $pcols[0] => { -in => $rs->as_query } },
1582 return $self->_multipk_update_delete (@_);
1586 # ANSI SQL does not provide a reliable way to perform a multicol-PK
1587 # resultset update/delete involving subqueries. So by default resort
1588 # to simple (and inefficient) delete_all style per-row opearations,
1589 # while allowing specific storages to override this with a faster
1592 sub _multipk_update_delete {
1593 return shift->_per_row_update_delete (@_);
1596 # This is the default loop used to delete/update rows for multi PK
1597 # resultsets, and used by mysql exclusively (because it can't do anything
1600 # We do not use $row->$op style queries, because resultset update/delete
1601 # is not expected to cascade (this is what delete_all/update_all is for).
1603 # There should be no race conditions as the entire operation is rolled
1606 sub _per_row_update_delete {
1608 my ($rs, $op, $values) = @_;
1610 my $rsrc = $rs->result_source;
1611 my @pcols = $rsrc->primary_columns;
1613 my $guard = $self->txn_scope_guard;
1615 # emulate the return value of $sth->execute for non-selects
1616 my $row_cnt = '0E0';
1618 my $subrs_cur = $rs->cursor;
1619 while (my @pks = $subrs_cur->next) {
1622 for my $i (0.. $#pcols) {
1623 $cond->{$pcols[$i]} = $pks[$i];
1628 $op eq 'update' ? $values : (),
1643 # localization is neccessary as
1644 # 1) there is no infrastructure to pass this around before SQLA2
1645 # 2) _select_args sets it and _prep_for_execute consumes it
1646 my $sql_maker = $self->sql_maker;
1647 local $sql_maker->{_dbic_rs_attrs};
1649 return $self->_execute($self->_select_args(@_));
1652 sub _select_args_to_query {
1655 # localization is neccessary as
1656 # 1) there is no infrastructure to pass this around before SQLA2
1657 # 2) _select_args sets it and _prep_for_execute consumes it
1658 my $sql_maker = $self->sql_maker;
1659 local $sql_maker->{_dbic_rs_attrs};
1661 # my ($op, $bind, $ident, $bind_attrs, $select, $cond, $order, $rows, $offset)
1662 # = $self->_select_args($ident, $select, $cond, $attrs);
1663 my ($op, $bind, $ident, $bind_attrs, @args) =
1664 $self->_select_args(@_);
1666 # my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, [ $select, $cond, $order, $rows, $offset ]);
1667 my ($sql, $prepared_bind) = $self->_prep_for_execute($op, $bind, $ident, \@args);
1668 $prepared_bind ||= [];
1671 ? ($sql, $prepared_bind, $bind_attrs)
1672 : \[ "($sql)", @$prepared_bind ]
1677 my ($self, $ident, $select, $where, $attrs) = @_;
1679 my ($alias2source, $rs_alias) = $self->_resolve_ident_sources ($ident);
1681 my $sql_maker = $self->sql_maker;
1682 $sql_maker->{_dbic_rs_attrs} = {
1688 ? ( _source_handle => $alias2source->{$rs_alias}->handle )
1693 # calculate bind_attrs before possible $ident mangling
1694 my $bind_attrs = {};
1695 for my $alias (keys %$alias2source) {
1696 my $bindtypes = $self->source_bind_attributes ($alias2source->{$alias}) || {};
1697 for my $col (keys %$bindtypes) {
1699 my $fqcn = join ('.', $alias, $col);
1700 $bind_attrs->{$fqcn} = $bindtypes->{$col} if $bindtypes->{$col};
1702 # Unqialified column names are nice, but at the same time can be
1703 # rather ambiguous. What we do here is basically go along with
1704 # the loop, adding an unqualified column slot to $bind_attrs,
1705 # alongside the fully qualified name. As soon as we encounter
1706 # another column by that name (which would imply another table)
1707 # we unset the unqualified slot and never add any info to it
1708 # to avoid erroneous type binding. If this happens the users
1709 # only choice will be to fully qualify his column name
1711 if (exists $bind_attrs->{$col}) {
1712 $bind_attrs->{$col} = {};
1715 $bind_attrs->{$col} = $bind_attrs->{$fqcn};
1722 $attrs->{software_limit}
1724 $sql_maker->_default_limit_syntax eq "GenericSubQ"
1726 $attrs->{software_limit} = 1;
1729 $self->throw_exception("rows attribute must be positive if present")
1730 if (defined($attrs->{rows}) && !($attrs->{rows} > 0));
1732 # MySQL actually recommends this approach. I cringe.
1733 $attrs->{rows} = 2**48 if not defined $attrs->{rows} and defined $attrs->{offset};
1738 # see if we need to tear the prefetch apart (either limited has_many or grouped prefetch)
1739 # otherwise delegate the limiting to the storage, unless software limit was requested
1741 ( $attrs->{rows} && keys %{$attrs->{collapse}} )
1743 ( $attrs->{group_by} && @{$attrs->{group_by}} &&
1744 $attrs->{_prefetch_select} && @{$attrs->{_prefetch_select}} )
1746 ($ident, $select, $where, $attrs)
1747 = $self->_adjust_select_args_for_complex_prefetch ($ident, $select, $where, $attrs);
1749 elsif (! $attrs->{software_limit} ) {
1750 push @limit, $attrs->{rows}, $attrs->{offset};
1754 # This would be the point to deflate anything found in $where
1755 # (and leave $attrs->{bind} intact). Problem is - inflators historically
1756 # expect a row object. And all we have is a resultsource (it is trivial
1757 # to extract deflator coderefs via $alias2source above).
1759 # I don't see a way forward other than changing the way deflators are
1760 # invoked, and that's just bad...
1764 { $attrs->{$_} ? ( $_ => $attrs->{$_} ) : () }
1765 (qw/order_by group_by having/ )
1768 return ('select', $attrs->{bind}, $ident, $bind_attrs, $select, $where, $order, @limit);
1772 # This is the code producing joined subqueries like:
1773 # SELECT me.*, other.* FROM ( SELECT me.* FROM ... ) JOIN other ON ...
1775 sub _adjust_select_args_for_complex_prefetch {
1776 my ($self, $from, $select, $where, $attrs) = @_;
1778 $self->throw_exception ('Nothing to prefetch... how did we get here?!')
1779 if not @{$attrs->{_prefetch_select}};
1781 $self->throw_exception ('Complex prefetches are not supported on resultsets with a custom from attribute')
1782 if (ref $from ne 'ARRAY' || ref $from->[0] ne 'HASH' || ref $from->[1] ne 'ARRAY');
1785 # generate inner/outer attribute lists, remove stuff that doesn't apply
1786 my $outer_attrs = { %$attrs };
1787 delete $outer_attrs->{$_} for qw/where bind rows offset group_by having/;
1789 my $inner_attrs = { %$attrs };
1790 delete $inner_attrs->{$_} for qw/for collapse _prefetch_select _collapse_order_by select as/;
1793 # bring over all non-collapse-induced order_by into the inner query (if any)
1794 # the outer one will have to keep them all
1795 delete $inner_attrs->{order_by};
1796 if (my $ord_cnt = @{$outer_attrs->{order_by}} - @{$outer_attrs->{_collapse_order_by}} ) {
1797 $inner_attrs->{order_by} = [
1798 @{$outer_attrs->{order_by}}[ 0 .. $ord_cnt - 1]
1803 # generate the inner/outer select lists
1804 # for inside we consider only stuff *not* brought in by the prefetch
1805 # on the outside we substitute any function for its alias
1806 my $outer_select = [ @$select ];
1807 my $inner_select = [];
1808 for my $i (0 .. ( @$outer_select - @{$outer_attrs->{_prefetch_select}} - 1) ) {
1809 my $sel = $outer_select->[$i];
1811 if (ref $sel eq 'HASH' ) {
1812 $sel->{-as} ||= $attrs->{as}[$i];
1813 $outer_select->[$i] = join ('.', $attrs->{alias}, ($sel->{-as} || "inner_column_$i") );
1816 push @$inner_select, $sel;
1819 # normalize a copy of $from, so it will be easier to work with further
1820 # down (i.e. promote the initial hashref to an AoH)
1822 $from->[0] = [ $from->[0] ];
1823 my %original_join_info = map { $_->[0]{-alias} => $_->[0] } (@$from);
1826 # decide which parts of the join will remain in either part of
1827 # the outer/inner query
1829 # First we compose a list of which aliases are used in restrictions
1830 # (i.e. conditions/order/grouping/etc). Since we do not have
1831 # introspectable SQLA, we fall back to ugly scanning of raw SQL for
1832 # WHERE, and for pieces of ORDER BY in order to determine which aliases
1833 # need to appear in the resulting sql.
1834 # It may not be very efficient, but it's a reasonable stop-gap
1835 # Also unqualified column names will not be considered, but more often
1836 # than not this is actually ok
1838 # In the same loop we enumerate part of the selection aliases, as
1839 # it requires the same sqla hack for the time being
1840 my ($restrict_aliases, $select_aliases, $prefetch_aliases);
1842 # produce stuff unquoted, so it can be scanned
1843 my $sql_maker = $self->sql_maker;
1844 local $sql_maker->{quote_char};
1845 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
1848 my $non_prefetch_select_sql = $sql_maker->_recurse_fields ($inner_select);
1849 my $prefetch_select_sql = $sql_maker->_recurse_fields ($outer_attrs->{_prefetch_select});
1850 my $where_sql = $sql_maker->where ($where);
1851 my $group_by_sql = $sql_maker->_order_by({
1852 map { $_ => $inner_attrs->{$_} } qw/group_by having/
1854 my @non_prefetch_order_by_chunks = (map
1855 { ref $_ ? $_->[0] : $_ }
1856 $sql_maker->_order_by_chunks ($inner_attrs->{order_by})
1860 for my $alias (keys %original_join_info) {
1861 my $seen_re = qr/\b $alias $sep/x;
1863 for my $piece ($where_sql, $group_by_sql, @non_prefetch_order_by_chunks ) {
1864 if ($piece =~ $seen_re) {
1865 $restrict_aliases->{$alias} = 1;
1869 if ($non_prefetch_select_sql =~ $seen_re) {
1870 $select_aliases->{$alias} = 1;
1873 if ($prefetch_select_sql =~ $seen_re) {
1874 $prefetch_aliases->{$alias} = 1;
1880 # Add any non-left joins to the restriction list (such joins are indeed restrictions)
1881 for my $j (values %original_join_info) {
1882 my $alias = $j->{-alias} or next;
1883 $restrict_aliases->{$alias} = 1 if (
1884 (not $j->{-join_type})
1886 ($j->{-join_type} !~ /^left (?: \s+ outer)? $/xi)
1890 # mark all join parents as mentioned
1891 # (e.g. join => { cds => 'tracks' } - tracks will need to bring cds too )
1892 for my $collection ($restrict_aliases, $select_aliases) {
1893 for my $alias (keys %$collection) {
1894 $collection->{$_} = 1
1895 for (@{ $original_join_info{$alias}{-join_path} || [] });
1899 # construct the inner $from for the subquery
1900 my %inner_joins = (map { %{$_ || {}} } ($restrict_aliases, $select_aliases) );
1902 for my $j (@$from) {
1903 push @inner_from, $j if $inner_joins{$j->[0]{-alias}};
1906 # if a multi-type join was needed in the subquery ("multi" is indicated by
1907 # presence in {collapse}) - add a group_by to simulate the collapse in the subq
1908 unless ($inner_attrs->{group_by}) {
1909 for my $alias (keys %inner_joins) {
1911 # the dot comes from some weirdness in collapse
1912 # remove after the rewrite
1913 if ($attrs->{collapse}{".$alias"}) {
1914 $inner_attrs->{group_by} ||= $inner_select;
1920 # demote the inner_from head
1921 $inner_from[0] = $inner_from[0][0];
1923 # generate the subquery
1924 my $subq = $self->_select_args_to_query (
1931 my $subq_joinspec = {
1932 -alias => $attrs->{alias},
1933 -source_handle => $inner_from[0]{-source_handle},
1934 $attrs->{alias} => $subq,
1937 # Generate the outer from - this is relatively easy (really just replace
1938 # the join slot with the subquery), with a major caveat - we can not
1939 # join anything that is non-selecting (not part of the prefetch), but at
1940 # the same time is a multi-type relationship, as it will explode the result.
1942 # There are two possibilities here
1943 # - either the join is non-restricting, in which case we simply throw it away
1944 # - it is part of the restrictions, in which case we need to collapse the outer
1945 # result by tackling yet another group_by to the outside of the query
1947 # so first generate the outer_from, up to the substitution point
1949 while (my $j = shift @$from) {
1950 if ($j->[0]{-alias} eq $attrs->{alias}) { # time to swap
1955 last; # we'll take care of what's left in $from below
1958 push @outer_from, $j;
1962 # see what's left - throw away if not selecting/restricting
1963 # also throw in a group_by if restricting to guard against
1964 # cross-join explosions
1966 while (my $j = shift @$from) {
1967 my $alias = $j->[0]{-alias};
1969 if ($select_aliases->{$alias} || $prefetch_aliases->{$alias}) {
1970 push @outer_from, $j;
1972 elsif ($restrict_aliases->{$alias}) {
1973 push @outer_from, $j;
1975 # FIXME - this should be obviated by SQLA2, as I'll be able to
1976 # have restrict_inner and restrict_outer... or something to that
1977 # effect... I think...
1979 # FIXME2 - I can't find a clean way to determine if a particular join
1980 # is a multi - instead I am just treating everything as a potential
1981 # explosive join (ribasushi)
1983 # if (my $handle = $j->[0]{-source_handle}) {
1984 # my $rsrc = $handle->resolve;
1985 # ... need to bail out of the following if this is not a multi,
1986 # as it will be much easier on the db ...
1988 $outer_attrs->{group_by} ||= $outer_select;
1993 # demote the outer_from head
1994 $outer_from[0] = $outer_from[0][0];
1996 # This is totally horrific - the $where ends up in both the inner and outer query
1997 # Unfortunately not much can be done until SQLA2 introspection arrives, and even
1998 # then if where conditions apply to the *right* side of the prefetch, you may have
1999 # to both filter the inner select (e.g. to apply a limit) and then have to re-filter
2000 # the outer select to exclude joins you didin't want in the first place
2002 # OTOH it can be seen as a plus: <ash> (notes that this query would make a DBA cry ;)
2003 return (\@outer_from, $outer_select, $where, $outer_attrs);
2006 sub _resolve_ident_sources {
2007 my ($self, $ident) = @_;
2009 my $alias2source = {};
2012 # the reason this is so contrived is that $ident may be a {from}
2013 # structure, specifying multiple tables to join
2014 if ( Scalar::Util::blessed($ident) && $ident->isa("DBIx::Class::ResultSource") ) {
2015 # this is compat mode for insert/update/delete which do not deal with aliases
2016 $alias2source->{me} = $ident;
2019 elsif (ref $ident eq 'ARRAY') {
2023 if (ref $_ eq 'HASH') {
2025 $rs_alias = $tabinfo->{-alias};
2027 if (ref $_ eq 'ARRAY' and ref $_->[0] eq 'HASH') {
2031 $alias2source->{$tabinfo->{-alias}} = $tabinfo->{-source_handle}->resolve
2032 if ($tabinfo->{-source_handle});
2036 return ($alias2source, $rs_alias);
2039 # Takes $ident, \@column_names
2041 # returns { $column_name => \%column_info, ... }
2042 # also note: this adds -result_source => $rsrc to the column info
2045 # my $col_sources = $self->_resolve_column_info($ident, @column_names);
2046 sub _resolve_column_info {
2047 my ($self, $ident, $colnames) = @_;
2048 my ($alias2src, $root_alias) = $self->_resolve_ident_sources($ident);
2050 my $sep = $self->_sql_maker_opts->{name_sep} || '.';
2053 my (%return, %seen_cols);
2055 # compile a global list of column names, to be able to properly
2056 # disambiguate unqualified column names (if at all possible)
2057 for my $alias (keys %$alias2src) {
2058 my $rsrc = $alias2src->{$alias};
2059 for my $colname ($rsrc->columns) {
2060 push @{$seen_cols{$colname}}, $alias;
2065 foreach my $col (@$colnames) {
2066 my ($alias, $colname) = $col =~ m/^ (?: ([^$sep]+) $sep)? (.+) $/x;
2069 # see if the column was seen exactly once (so we know which rsrc it came from)
2070 if ($seen_cols{$colname} and @{$seen_cols{$colname}} == 1) {
2071 $alias = $seen_cols{$colname}[0];
2078 my $rsrc = $alias2src->{$alias};
2079 $return{$col} = $rsrc && {
2080 %{$rsrc->column_info($colname)},
2081 -result_source => $rsrc,
2082 -source_alias => $alias,
2089 # Returns a counting SELECT for a simple count
2090 # query. Abstracted so that a storage could override
2091 # this to { count => 'firstcol' } or whatever makes
2092 # sense as a performance optimization
2094 #my ($self, $source, $rs_attrs) = @_;
2095 return { count => '*' };
2098 # Returns a SELECT which will end up in the subselect
2099 # There may or may not be a group_by, as the subquery
2100 # might have been called to accomodate a limit
2102 # Most databases would be happy with whatever ends up
2103 # here, but some choke in various ways.
2105 sub _subq_count_select {
2106 my ($self, $source, $rs_attrs) = @_;
2107 return $rs_attrs->{group_by} if $rs_attrs->{group_by};
2109 my @pcols = map { join '.', $rs_attrs->{alias}, $_ } ($source->primary_columns);
2110 return @pcols ? \@pcols : [ 1 ];
2113 sub source_bind_attributes {
2114 my ($self, $source) = @_;
2116 my $bind_attributes;
2117 foreach my $column ($source->columns) {
2119 my $data_type = $source->column_info($column)->{data_type} || '';
2120 $bind_attributes->{$column} = $self->bind_attribute_by_data_type($data_type)
2124 return $bind_attributes;
2131 =item Arguments: $ident, $select, $condition, $attrs
2135 Handle a SQL select statement.
2141 my ($ident, $select, $condition, $attrs) = @_;
2142 return $self->cursor_class->new($self, \@_, $attrs);
2147 my ($rv, $sth, @bind) = $self->_select(@_);
2148 my @row = $sth->fetchrow_array;
2149 my @nextrow = $sth->fetchrow_array if @row;
2150 if(@row && @nextrow) {
2151 carp "Query returned more than one row. SQL that returns multiple rows is DEPRECATED for ->find and ->single";
2153 # Need to call finish() to work round broken DBDs
2162 =item Arguments: $sql
2166 Returns a L<DBI> sth (statement handle) for the supplied SQL.
2171 my ($self, $dbh, $sql) = @_;
2173 # 3 is the if_active parameter which avoids active sth re-use
2174 my $sth = $self->disable_sth_caching
2175 ? $dbh->prepare($sql)
2176 : $dbh->prepare_cached($sql, {}, 3);
2178 # XXX You would think RaiseError would make this impossible,
2179 # but apparently that's not true :(
2180 $self->throw_exception($dbh->errstr) if !$sth;
2186 my ($self, $sql) = @_;
2187 $self->dbh_do('_dbh_sth', $sql); # retry over disconnects
2190 sub _dbh_columns_info_for {
2191 my ($self, $dbh, $table) = @_;
2193 if ($dbh->can('column_info')) {
2196 my ($schema,$tab) = $table =~ /^(.+?)\.(.+)$/ ? ($1,$2) : (undef,$table);
2197 my $sth = $dbh->column_info( undef,$schema, $tab, '%' );
2199 while ( my $info = $sth->fetchrow_hashref() ){
2201 $column_info{data_type} = $info->{TYPE_NAME};
2202 $column_info{size} = $info->{COLUMN_SIZE};
2203 $column_info{is_nullable} = $info->{NULLABLE} ? 1 : 0;
2204 $column_info{default_value} = $info->{COLUMN_DEF};
2205 my $col_name = $info->{COLUMN_NAME};
2206 $col_name =~ s/^\"(.*)\"$/$1/;
2208 $result{$col_name} = \%column_info;
2211 return \%result if !$@ && scalar keys %result;
2215 my $sth = $dbh->prepare($self->sql_maker->select($table, undef, \'1 = 0'));
2217 my @columns = @{$sth->{NAME_lc}};
2218 for my $i ( 0 .. $#columns ){
2220 $column_info{data_type} = $sth->{TYPE}->[$i];
2221 $column_info{size} = $sth->{PRECISION}->[$i];
2222 $column_info{is_nullable} = $sth->{NULLABLE}->[$i] ? 1 : 0;
2224 if ($column_info{data_type} =~ m/^(.*?)\((.*?)\)$/) {
2225 $column_info{data_type} = $1;
2226 $column_info{size} = $2;
2229 $result{$columns[$i]} = \%column_info;
2233 foreach my $col (keys %result) {
2234 my $colinfo = $result{$col};
2235 my $type_num = $colinfo->{data_type};
2237 if(defined $type_num && $dbh->can('type_info')) {
2238 my $type_info = $dbh->type_info($type_num);
2239 $type_name = $type_info->{TYPE_NAME} if $type_info;
2240 $colinfo->{data_type} = $type_name if $type_name;
2247 sub columns_info_for {
2248 my ($self, $table) = @_;
2249 $self->_dbh_columns_info_for ($self->_get_dbh, $table);
2252 =head2 last_insert_id
2254 Return the row id of the last insert.
2258 sub _dbh_last_insert_id {
2259 # All Storage's need to register their own _dbh_last_insert_id
2260 # the old SQLite-based method was highly inappropriate
2263 my $class = ref $self;
2264 $self->throw_exception (<<EOE);
2266 No _dbh_last_insert_id() method found in $class.
2267 Since the method of obtaining the autoincrement id of the last insert
2268 operation varies greatly between different databases, this method must be
2269 individually implemented for every storage class.
2273 sub last_insert_id {
2275 $self->_dbh_last_insert_id ($self->_dbh, @_);
2278 =head2 _native_data_type
2282 =item Arguments: $type_name
2286 This API is B<EXPERIMENTAL>, will almost definitely change in the future, and
2287 currently only used by L<::AutoCast|DBIx::Class::Storage::DBI::AutoCast> and
2288 L<::Sybase|DBIx::Class::Storage::DBI::Sybase>.
2290 The default implementation returns C<undef>, implement in your Storage driver if
2291 you need this functionality.
2293 Should map types from other databases to the native RDBMS type, for example
2294 C<VARCHAR2> to C<VARCHAR>.
2296 Types with modifiers should map to the underlying data type. For example,
2297 C<INTEGER AUTO_INCREMENT> should become C<INTEGER>.
2299 Composite types should map to the container type, for example
2300 C<ENUM(foo,bar,baz)> becomes C<ENUM>.
2304 sub _native_data_type {
2305 #my ($self, $data_type) = @_;
2309 # Check if placeholders are supported at all
2310 sub _placeholders_supported {
2312 my $dbh = $self->_get_dbh;
2314 # some drivers provide a $dbh attribute (e.g. Sybase and $dbh->{syb_dynamic_supported})
2315 # but it is inaccurate more often than not
2317 local $dbh->{PrintError} = 0;
2318 local $dbh->{RaiseError} = 1;
2319 $dbh->do('select ?', {}, 1);
2324 # Check if placeholders bound to non-string types throw exceptions
2326 sub _typeless_placeholders_supported {
2328 my $dbh = $self->_get_dbh;
2331 local $dbh->{PrintError} = 0;
2332 local $dbh->{RaiseError} = 1;
2333 # this specifically tests a bind that is NOT a string
2334 $dbh->do('select 1 where 1 = ?', {}, 1);
2341 Returns the database driver name.
2348 if (not $self->_driver_determined) {
2349 $self->_determine_driver;
2350 goto $self->can ('sqlt_type');
2353 $self->_get_dbh->{Driver}->{Name};
2356 =head2 bind_attribute_by_data_type
2358 Given a datatype from column info, returns a database specific bind
2359 attribute for C<< $dbh->bind_param($val,$attribute) >> or nothing if we will
2360 let the database planner just handle it.
2362 Generally only needed for special case column types, like bytea in postgres.
2366 sub bind_attribute_by_data_type {
2370 =head2 is_datatype_numeric
2372 Given a datatype from column_info, returns a boolean value indicating if
2373 the current RDBMS considers it a numeric value. This controls how
2374 L<DBIx::Class::Row/set_column> decides whether to mark the column as
2375 dirty - when the datatype is deemed numeric a C<< != >> comparison will
2376 be performed instead of the usual C<eq>.
2380 sub is_datatype_numeric {
2381 my ($self, $dt) = @_;
2383 return 0 unless $dt;
2385 return $dt =~ /^ (?:
2386 numeric | int(?:eger)? | (?:tiny|small|medium|big)int | dec(?:imal)? | real | float | double (?: \s+ precision)? | (?:big)?serial
2391 =head2 create_ddl_dir (EXPERIMENTAL)
2395 =item Arguments: $schema \@databases, $version, $directory, $preversion, \%sqlt_args
2399 Creates a SQL file based on the Schema, for each of the specified
2400 database engines in C<\@databases> in the given directory.
2401 (note: specify L<SQL::Translator> names, not L<DBI> driver names).
2403 Given a previous version number, this will also create a file containing
2404 the ALTER TABLE statements to transform the previous schema into the
2405 current one. Note that these statements may contain C<DROP TABLE> or
2406 C<DROP COLUMN> statements that can potentially destroy data.
2408 The file names are created using the C<ddl_filename> method below, please
2409 override this method in your schema if you would like a different file
2410 name format. For the ALTER file, the same format is used, replacing
2411 $version in the name with "$preversion-$version".
2413 See L<SQL::Translator/METHODS> for a list of values for C<\%sqlt_args>.
2414 The most common value for this would be C<< { add_drop_table => 1 } >>
2415 to have the SQL produced include a C<DROP TABLE> statement for each table
2416 created. For quoting purposes supply C<quote_table_names> and
2417 C<quote_field_names>.
2419 If no arguments are passed, then the following default values are assumed:
2423 =item databases - ['MySQL', 'SQLite', 'PostgreSQL']
2425 =item version - $schema->schema_version
2427 =item directory - './'
2429 =item preversion - <none>
2433 By default, C<\%sqlt_args> will have
2435 { add_drop_table => 1, ignore_constraint_names => 1, ignore_index_names => 1 }
2437 merged with the hash passed in. To disable any of those features, pass in a
2438 hashref like the following
2440 { ignore_constraint_names => 0, # ... other options }
2443 Note that this feature is currently EXPERIMENTAL and may not work correctly
2444 across all databases, or fully handle complex relationships.
2446 WARNING: Please check all SQL files created, before applying them.
2450 sub create_ddl_dir {
2451 my ($self, $schema, $databases, $version, $dir, $preversion, $sqltargs) = @_;
2453 if(!$dir || !-d $dir) {
2454 carp "No directory given, using ./\n";
2457 $databases ||= ['MySQL', 'SQLite', 'PostgreSQL'];
2458 $databases = [ $databases ] if(ref($databases) ne 'ARRAY');
2460 my $schema_version = $schema->schema_version || '1.x';
2461 $version ||= $schema_version;
2464 add_drop_table => 1,
2465 ignore_constraint_names => 1,
2466 ignore_index_names => 1,
2470 $self->throw_exception("Can't create a ddl file without SQL::Translator: " . $self->_sqlt_version_error)
2471 if !$self->_sqlt_version_ok;
2473 my $sqlt = SQL::Translator->new( $sqltargs );
2475 $sqlt->parser('SQL::Translator::Parser::DBIx::Class');
2476 my $sqlt_schema = $sqlt->translate({ data => $schema })
2477 or $self->throw_exception ($sqlt->error);
2479 foreach my $db (@$databases) {
2481 $sqlt->{schema} = $sqlt_schema;
2482 $sqlt->producer($db);
2485 my $filename = $schema->ddl_filename($db, $version, $dir);
2486 if (-e $filename && ($version eq $schema_version )) {
2487 # if we are dumping the current version, overwrite the DDL
2488 carp "Overwriting existing DDL file - $filename";
2492 my $output = $sqlt->translate;
2494 carp("Failed to translate to $db, skipping. (" . $sqlt->error . ")");
2497 if(!open($file, ">$filename")) {
2498 $self->throw_exception("Can't open $filename for writing ($!)");
2501 print $file $output;
2504 next unless ($preversion);
2506 require SQL::Translator::Diff;
2508 my $prefilename = $schema->ddl_filename($db, $preversion, $dir);
2509 if(!-e $prefilename) {
2510 carp("No previous schema file found ($prefilename)");
2514 my $difffile = $schema->ddl_filename($db, $version, $dir, $preversion);
2516 carp("Overwriting existing diff file - $difffile");
2522 my $t = SQL::Translator->new($sqltargs);
2527 or $self->throw_exception ($t->error);
2529 my $out = $t->translate( $prefilename )
2530 or $self->throw_exception ($t->error);
2532 $source_schema = $t->schema;
2534 $source_schema->name( $prefilename )
2535 unless ( $source_schema->name );
2538 # The "new" style of producers have sane normalization and can support
2539 # diffing a SQL file against a DBIC->SQLT schema. Old style ones don't
2540 # And we have to diff parsed SQL against parsed SQL.
2541 my $dest_schema = $sqlt_schema;
2543 unless ( "SQL::Translator::Producer::$db"->can('preprocess_schema') ) {
2544 my $t = SQL::Translator->new($sqltargs);
2549 or $self->throw_exception ($t->error);
2551 my $out = $t->translate( $filename )
2552 or $self->throw_exception ($t->error);
2554 $dest_schema = $t->schema;
2556 $dest_schema->name( $filename )
2557 unless $dest_schema->name;
2560 my $diff = SQL::Translator::Diff::schema_diff($source_schema, $db,
2564 if(!open $file, ">$difffile") {
2565 $self->throw_exception("Can't write to $difffile ($!)");
2573 =head2 deployment_statements
2577 =item Arguments: $schema, $type, $version, $directory, $sqlt_args
2581 Returns the statements used by L</deploy> and L<DBIx::Class::Schema/deploy>.
2583 The L<SQL::Translator> (not L<DBI>) database driver name can be explicitly
2584 provided in C<$type>, otherwise the result of L</sqlt_type> is used as default.
2586 C<$directory> is used to return statements from files in a previously created
2587 L</create_ddl_dir> directory and is optional. The filenames are constructed
2588 from L<DBIx::Class::Schema/ddl_filename>, the schema name and the C<$version>.
2590 If no C<$directory> is specified then the statements are constructed on the
2591 fly using L<SQL::Translator> and C<$version> is ignored.
2593 See L<SQL::Translator/METHODS> for a list of values for C<$sqlt_args>.
2597 sub deployment_statements {
2598 my ($self, $schema, $type, $version, $dir, $sqltargs) = @_;
2599 $type ||= $self->sqlt_type;
2600 $version ||= $schema->schema_version || '1.x';
2602 my $filename = $schema->ddl_filename($type, $version, $dir);
2606 open($file, "<$filename")
2607 or $self->throw_exception("Can't open $filename ($!)");
2610 return join('', @rows);
2613 $self->throw_exception("Can't deploy without either SQL::Translator or a ddl_dir: " . $self->_sqlt_version_error )
2614 if !$self->_sqlt_version_ok;
2616 # sources needs to be a parser arg, but for simplicty allow at top level
2618 $sqltargs->{parser_args}{sources} = delete $sqltargs->{sources}
2619 if exists $sqltargs->{sources};
2621 my $tr = SQL::Translator->new(
2622 producer => "SQL::Translator::Producer::${type}",
2624 parser => 'SQL::Translator::Parser::DBIx::Class',
2627 return $tr->translate;
2631 my ($self, $schema, $type, $sqltargs, $dir) = @_;
2634 return if($line =~ /^--/);
2636 # next if($line =~ /^DROP/m);
2637 return if($line =~ /^BEGIN TRANSACTION/m);
2638 return if($line =~ /^COMMIT/m);
2639 return if $line =~ /^\s+$/; # skip whitespace only
2640 $self->_query_start($line);
2642 # do a dbh_do cycle here, as we need some error checking in
2643 # place (even though we will ignore errors)
2644 $self->dbh_do (sub { $_[1]->do($line) });
2647 carp qq{$@ (running "${line}")};
2649 $self->_query_end($line);
2651 my @statements = $self->deployment_statements($schema, $type, undef, $dir, { %{ $sqltargs || {} }, no_comments => 1 } );
2652 if (@statements > 1) {
2653 foreach my $statement (@statements) {
2654 $deploy->( $statement );
2657 elsif (@statements == 1) {
2658 foreach my $line ( split(";\n", $statements[0])) {
2664 =head2 datetime_parser
2666 Returns the datetime parser class
2670 sub datetime_parser {
2672 return $self->{datetime_parser} ||= do {
2673 $self->build_datetime_parser(@_);
2677 =head2 datetime_parser_type
2679 Defines (returns) the datetime parser class - currently hardwired to
2680 L<DateTime::Format::MySQL>
2684 sub datetime_parser_type { "DateTime::Format::MySQL"; }
2686 =head2 build_datetime_parser
2688 See L</datetime_parser>
2692 sub build_datetime_parser {
2693 if (not $_[0]->_driver_determined) {
2694 $_[0]->_determine_driver;
2695 goto $_[0]->can('build_datetime_parser');
2699 my $type = $self->datetime_parser_type(@_);
2700 $self->ensure_class_loaded ($type);
2705 =head2 is_replicating
2707 A boolean that reports if a particular L<DBIx::Class::Storage::DBI> is set to
2708 replicate from a master database. Default is undef, which is the result
2709 returned by databases that don't support replication.
2713 sub is_replicating {
2718 =head2 lag_behind_master
2720 Returns a number that represents a certain amount of lag behind a master db
2721 when a given storage is replicating. The number is database dependent, but
2722 starts at zero and increases with the amount of lag. Default in undef
2726 sub lag_behind_master {
2730 # SQLT version handling
2732 my $_sqlt_version_ok; # private
2733 my $_sqlt_version_error; # private
2735 sub _sqlt_version_ok {
2736 if (!defined $_sqlt_version_ok) {
2737 eval "use SQL::Translator $minimum_sqlt_version";
2739 $_sqlt_version_ok = 0;
2740 $_sqlt_version_error = $@;
2743 $_sqlt_version_ok = 1;
2746 return $_sqlt_version_ok;
2749 sub _sqlt_version_error {
2750 shift->_sqlt_version_ok unless defined $_sqlt_version_ok;
2751 return $_sqlt_version_error;
2754 sub _sqlt_minimum_version { $minimum_sqlt_version };
2760 $self->_verify_pid if $self->_dbh;
2762 # some databases need this to stop spewing warnings
2763 if (my $dbh = $self->_dbh) {
2765 eval { $dbh->disconnect };
2775 =head2 DBIx::Class and AutoCommit
2777 DBIx::Class can do some wonderful magic with handling exceptions,
2778 disconnections, and transactions when you use C<< AutoCommit => 1 >>
2779 (the default) combined with C<txn_do> for transaction support.
2781 If you set C<< AutoCommit => 0 >> in your connect info, then you are always
2782 in an assumed transaction between commits, and you're telling us you'd
2783 like to manage that manually. A lot of the magic protections offered by
2784 this module will go away. We can't protect you from exceptions due to database
2785 disconnects because we don't know anything about how to restart your
2786 transactions. You're on your own for handling all sorts of exceptional
2787 cases if you choose the C<< AutoCommit => 0 >> path, just as you would
2793 Matt S. Trout <mst@shadowcatsystems.co.uk>
2795 Andy Grundman <andy@hybridized.org>
2799 You may distribute this code under the same terms as Perl itself.