Commit | Line | Data |
057db5ce |
1 | package DBIx::Class::Storage::DBI::Sybase::ASE; |
2 | |
3 | use strict; |
4 | use warnings; |
5 | |
6 | use base qw/ |
7 | DBIx::Class::Storage::DBI::Sybase |
8 | DBIx::Class::Storage::DBI::AutoCast |
9 | /; |
10 | use mro 'c3'; |
11 | use Carp::Clan qw/^DBIx::Class/; |
6298a324 |
12 | use Scalar::Util 'blessed'; |
13 | use List::Util 'first'; |
057db5ce |
14 | use Sub::Name(); |
6298a324 |
15 | use Data::Dumper::Concise 'Dumper'; |
ed7ab0f4 |
16 | use Try::Tiny; |
fd323bf1 |
17 | use namespace::clean; |
057db5ce |
18 | |
19 | __PACKAGE__->mk_group_accessors('simple' => |
20 | qw/_identity _blob_log_on_update _writer_storage _is_extra_storage |
21 | _bulk_storage _is_bulk_storage _began_bulk_work |
22 | _bulk_disabled_due_to_coderef_connect_info_warned |
23 | _identity_method/ |
24 | ); |
25 | |
26 | my @also_proxy_to_extra_storages = qw/ |
27 | connect_call_set_auto_cast auto_cast connect_call_blob_setup |
28 | connect_call_datetime_setup |
29 | |
30 | disconnect _connect_info _sql_maker _sql_maker_opts disable_sth_caching |
31 | auto_savepoint unsafe cursor_class debug debugobj schema |
32 | /; |
33 | |
34 | =head1 NAME |
35 | |
36 | DBIx::Class::Storage::DBI::Sybase::ASE - Sybase ASE SQL Server support for |
37 | DBIx::Class |
38 | |
39 | =head1 SYNOPSIS |
40 | |
41 | This subclass supports L<DBD::Sybase> for real (non-Microsoft) Sybase databases. |
42 | |
43 | =head1 DESCRIPTION |
44 | |
45 | If your version of Sybase does not support placeholders, then your storage will |
46 | be reblessed to L<DBIx::Class::Storage::DBI::Sybase::ASE::NoBindVars>. |
47 | You can also enable that driver explicitly, see the documentation for more |
48 | details. |
49 | |
50 | With this driver there is unfortunately no way to get the C<last_insert_id> |
51 | without doing a C<SELECT MAX(col)>. This is done safely in a transaction |
52 | (locking the table.) See L</INSERTS WITH PLACEHOLDERS>. |
53 | |
8384a713 |
54 | A recommended L<connect_info|DBIx::Class::Storage::DBI/connect_info> setting: |
057db5ce |
55 | |
56 | on_connect_call => [['datetime_setup'], ['blob_setup', log_on_update => 0]] |
57 | |
58 | =head1 METHODS |
59 | |
60 | =cut |
61 | |
62 | sub _rebless { |
63 | my $self = shift; |
64 | |
65 | my $no_bind_vars = __PACKAGE__ . '::NoBindVars'; |
66 | |
67 | if ($self->using_freetds) { |
68 | carp <<'EOF' unless $ENV{DBIC_SYBASE_FREETDS_NOWARN}; |
69 | |
70 | You are using FreeTDS with Sybase. |
71 | |
72 | We will do our best to support this configuration, but please consider this |
73 | support experimental. |
74 | |
75 | TEXT/IMAGE columns will definitely not work. |
76 | |
77 | You are encouraged to recompile DBD::Sybase with the Sybase Open Client libraries |
78 | instead. |
79 | |
80 | See perldoc DBIx::Class::Storage::DBI::Sybase::ASE for more details. |
81 | |
82 | To turn off this warning set the DBIC_SYBASE_FREETDS_NOWARN environment |
83 | variable. |
84 | EOF |
85 | |
86 | if (not $self->_typeless_placeholders_supported) { |
87 | if ($self->_placeholders_supported) { |
88 | $self->auto_cast(1); |
89 | } |
90 | else { |
91 | $self->ensure_class_loaded($no_bind_vars); |
92 | bless $self, $no_bind_vars; |
93 | $self->_rebless; |
94 | } |
95 | } |
96 | } |
97 | |
98 | elsif (not $self->_get_dbh->{syb_dynamic_supported}) { |
99 | # not necessarily FreeTDS, but no placeholders nevertheless |
100 | $self->ensure_class_loaded($no_bind_vars); |
101 | bless $self, $no_bind_vars; |
102 | $self->_rebless; |
103 | } |
104 | # this is highly unlikely, but we check just in case |
105 | elsif (not $self->_typeless_placeholders_supported) { |
106 | $self->auto_cast(1); |
107 | } |
108 | } |
109 | |
110 | sub _init { |
111 | my $self = shift; |
112 | $self->_set_max_connect(256); |
113 | |
114 | # create storage for insert/(update blob) transactions, |
115 | # unless this is that storage |
116 | return if $self->_is_extra_storage; |
117 | |
118 | my $writer_storage = (ref $self)->new; |
119 | |
120 | $writer_storage->_is_extra_storage(1); |
121 | $writer_storage->connect_info($self->connect_info); |
122 | $writer_storage->auto_cast($self->auto_cast); |
123 | |
124 | $self->_writer_storage($writer_storage); |
125 | |
126 | # create a bulk storage unless connect_info is a coderef |
127 | return if ref($self->_dbi_connect_info->[0]) eq 'CODE'; |
128 | |
129 | my $bulk_storage = (ref $self)->new; |
130 | |
131 | $bulk_storage->_is_extra_storage(1); |
132 | $bulk_storage->_is_bulk_storage(1); # for special ->disconnect acrobatics |
133 | $bulk_storage->connect_info($self->connect_info); |
134 | |
135 | # this is why |
136 | $bulk_storage->_dbi_connect_info->[0] .= ';bulkLogin=1'; |
137 | |
138 | $self->_bulk_storage($bulk_storage); |
139 | } |
140 | |
141 | for my $method (@also_proxy_to_extra_storages) { |
142 | no strict 'refs'; |
143 | no warnings 'redefine'; |
144 | |
145 | my $replaced = __PACKAGE__->can($method); |
146 | |
147 | *{$method} = Sub::Name::subname $method => sub { |
148 | my $self = shift; |
149 | $self->_writer_storage->$replaced(@_) if $self->_writer_storage; |
150 | $self->_bulk_storage->$replaced(@_) if $self->_bulk_storage; |
151 | return $self->$replaced(@_); |
152 | }; |
153 | } |
154 | |
3f528319 |
155 | sub _sql_maker_opts { |
156 | my ( $self, $opts ) = @_; |
157 | |
158 | if ( $opts ) { |
159 | $self->{_sql_maker_opts} = { %$opts }; |
160 | } |
161 | |
162 | return { limit_dialect => 'RowCountOrGenericSubQ', %{$self->{_sql_maker_opts}||{}} }; |
163 | } |
164 | |
057db5ce |
165 | sub disconnect { |
166 | my $self = shift; |
167 | |
168 | # Even though we call $sth->finish for uses off the bulk API, there's still an |
169 | # "active statement" warning on disconnect, which we throw away here. |
170 | # This is due to the bug described in insert_bulk. |
171 | # Currently a noop because 'prepare' is used instead of 'prepare_cached'. |
172 | local $SIG{__WARN__} = sub { |
173 | warn $_[0] unless $_[0] =~ /active statement/i; |
174 | } if $self->_is_bulk_storage; |
175 | |
176 | # so that next transaction gets a dbh |
177 | $self->_began_bulk_work(0) if $self->_is_bulk_storage; |
178 | |
179 | $self->next::method; |
180 | } |
181 | |
182 | # Set up session settings for Sybase databases for the connection. |
183 | # |
184 | # Make sure we have CHAINED mode turned on if AutoCommit is off in non-FreeTDS |
185 | # DBD::Sybase (since we don't know how DBD::Sybase was compiled.) If however |
186 | # we're using FreeTDS, CHAINED mode turns on an implicit transaction which we |
187 | # only want when AutoCommit is off. |
188 | # |
189 | # Also SET TEXTSIZE for FreeTDS because LongReadLen doesn't work. |
190 | sub _run_connection_actions { |
191 | my $self = shift; |
192 | |
193 | if ($self->_is_bulk_storage) { |
194 | # this should be cleared on every reconnect |
195 | $self->_began_bulk_work(0); |
196 | return; |
197 | } |
198 | |
199 | if (not $self->using_freetds) { |
200 | $self->_dbh->{syb_chained_txn} = 1; |
201 | } else { |
202 | # based on LongReadLen in connect_info |
203 | $self->set_textsize; |
204 | |
205 | if ($self->_dbh_autocommit) { |
206 | $self->_dbh->do('SET CHAINED OFF'); |
207 | } else { |
208 | $self->_dbh->do('SET CHAINED ON'); |
209 | } |
210 | } |
211 | |
212 | $self->next::method(@_); |
213 | } |
214 | |
215 | =head2 connect_call_blob_setup |
216 | |
217 | Used as: |
218 | |
219 | on_connect_call => [ [ 'blob_setup', log_on_update => 0 ] ] |
220 | |
221 | Does C<< $dbh->{syb_binary_images} = 1; >> to return C<IMAGE> data as raw binary |
222 | instead of as a hex string. |
223 | |
224 | Recommended. |
225 | |
226 | Also sets the C<log_on_update> value for blob write operations. The default is |
227 | C<1>, but C<0> is better if your database is configured for it. |
228 | |
229 | See |
230 | L<DBD::Sybase/Handling_IMAGE/TEXT_data_with_syb_ct_get_data()/syb_ct_send_data()>. |
231 | |
232 | =cut |
233 | |
234 | sub connect_call_blob_setup { |
235 | my $self = shift; |
236 | my %args = @_; |
237 | my $dbh = $self->_dbh; |
238 | $dbh->{syb_binary_images} = 1; |
239 | |
240 | $self->_blob_log_on_update($args{log_on_update}) |
241 | if exists $args{log_on_update}; |
242 | } |
243 | |
244 | sub _is_lob_type { |
245 | my $self = shift; |
246 | my $type = shift; |
247 | $type && $type =~ /(?:text|image|lob|bytea|binary|memo)/i; |
248 | } |
249 | |
250 | sub _is_lob_column { |
251 | my ($self, $source, $column) = @_; |
252 | |
253 | return $self->_is_lob_type($source->column_info($column)->{data_type}); |
254 | } |
255 | |
256 | sub _prep_for_execute { |
257 | my $self = shift; |
258 | my ($op, $extra_bind, $ident, $args) = @_; |
259 | |
260 | my ($sql, $bind) = $self->next::method (@_); |
261 | |
6298a324 |
262 | my $table = blessed $ident ? $ident->from : $ident; |
057db5ce |
263 | |
264 | my $bind_info = $self->_resolve_column_info( |
265 | $ident, [map $_->[0], @{$bind}] |
266 | ); |
6298a324 |
267 | my $bound_identity_col = |
268 | first { $bind_info->{$_}{is_auto_increment} } |
269 | keys %$bind_info |
057db5ce |
270 | ; |
6298a324 |
271 | my $identity_col = |
272 | blessed $ident && |
273 | first { $ident->column_info($_)->{is_auto_increment} } $ident->columns |
057db5ce |
274 | ; |
275 | |
276 | if (($op eq 'insert' && $bound_identity_col) || |
277 | ($op eq 'update' && exists $args->[0]{$identity_col})) { |
278 | $sql = join ("\n", |
279 | $self->_set_table_identity_sql($op => $table, 'on'), |
280 | $sql, |
281 | $self->_set_table_identity_sql($op => $table, 'off'), |
282 | ); |
283 | } |
284 | |
285 | if ($op eq 'insert' && (not $bound_identity_col) && $identity_col && |
286 | (not $self->{insert_bulk})) { |
287 | $sql = |
288 | "$sql\n" . |
289 | $self->_fetch_identity_sql($ident, $identity_col); |
290 | } |
291 | |
292 | return ($sql, $bind); |
293 | } |
294 | |
295 | sub _set_table_identity_sql { |
296 | my ($self, $op, $table, $on_off) = @_; |
297 | |
298 | return sprintf 'SET IDENTITY_%s %s %s', |
299 | uc($op), $self->sql_maker->_quote($table), uc($on_off); |
300 | } |
301 | |
302 | # Stolen from SQLT, with some modifications. This is a makeshift |
303 | # solution before a sane type-mapping library is available, thus |
304 | # the 'our' for easy overrides. |
305 | our %TYPE_MAPPING = ( |
306 | number => 'numeric', |
307 | money => 'money', |
308 | varchar => 'varchar', |
309 | varchar2 => 'varchar', |
310 | timestamp => 'datetime', |
311 | text => 'varchar', |
312 | real => 'double precision', |
313 | comment => 'text', |
314 | bit => 'bit', |
315 | tinyint => 'smallint', |
316 | float => 'double precision', |
317 | serial => 'numeric', |
318 | bigserial => 'numeric', |
319 | boolean => 'varchar', |
320 | long => 'varchar', |
321 | ); |
322 | |
323 | sub _native_data_type { |
324 | my ($self, $type) = @_; |
325 | |
326 | $type = lc $type; |
327 | $type =~ s/\s* identity//x; |
328 | |
329 | return uc($TYPE_MAPPING{$type} || $type); |
330 | } |
331 | |
332 | sub _fetch_identity_sql { |
333 | my ($self, $source, $col) = @_; |
334 | |
335 | return sprintf ("SELECT MAX(%s) FROM %s", |
336 | map { $self->sql_maker->_quote ($_) } ($col, $source->from) |
337 | ); |
338 | } |
339 | |
340 | sub _execute { |
341 | my $self = shift; |
342 | my ($op) = @_; |
343 | |
344 | my ($rv, $sth, @bind) = $self->dbh_do($self->can('_dbh_execute'), @_); |
345 | |
346 | if ($op eq 'insert') { |
347 | $self->_identity($sth->fetchrow_array); |
348 | $sth->finish; |
349 | } |
350 | |
351 | return wantarray ? ($rv, $sth, @bind) : $rv; |
352 | } |
353 | |
354 | sub last_insert_id { shift->_identity } |
355 | |
356 | # handles TEXT/IMAGE and transaction for last_insert_id |
357 | sub insert { |
358 | my $self = shift; |
359 | my ($source, $to_insert) = @_; |
360 | |
6298a324 |
361 | my $identity_col = |
362 | (first { $source->column_info($_)->{is_auto_increment} } $source->columns) |
363 | || ''; |
057db5ce |
364 | |
365 | # check for empty insert |
366 | # INSERT INTO foo DEFAULT VALUES -- does not work with Sybase |
6469dabf |
367 | # try to insert explicit 'DEFAULT's instead (except for identity, timestamp |
368 | # and computed columns) |
057db5ce |
369 | if (not %$to_insert) { |
370 | for my $col ($source->columns) { |
371 | next if $col eq $identity_col; |
6469dabf |
372 | |
373 | my $info = $source->column_info($col); |
374 | |
375 | next if ref $info->{default_value} eq 'SCALAR' |
376 | || (exists $info->{data_type} && (not defined $info->{data_type})); |
377 | |
378 | next if $info->{data_type} && $info->{data_type} =~ /^timestamp\z/i; |
379 | |
057db5ce |
380 | $to_insert->{$col} = \'DEFAULT'; |
381 | } |
382 | } |
383 | |
384 | my $blob_cols = $self->_remove_blob_cols($source, $to_insert); |
385 | |
386 | # do we need the horrific SELECT MAX(COL) hack? |
387 | my $dumb_last_insert_id = |
388 | $identity_col |
389 | && (not exists $to_insert->{$identity_col}) |
390 | && ($self->_identity_method||'') ne '@@IDENTITY'; |
391 | |
392 | my $next = $self->next::can; |
393 | |
394 | # we are already in a transaction, or there are no blobs |
395 | # and we don't need the PK - just (try to) do it |
396 | if ($self->{transaction_depth} |
397 | || (!$blob_cols && !$dumb_last_insert_id) |
398 | ) { |
399 | return $self->_insert ( |
400 | $next, $source, $to_insert, $blob_cols, $identity_col |
401 | ); |
402 | } |
403 | |
404 | # otherwise use the _writer_storage to do the insert+transaction on another |
405 | # connection |
406 | my $guard = $self->_writer_storage->txn_scope_guard; |
407 | |
408 | my $updated_cols = $self->_writer_storage->_insert ( |
409 | $next, $source, $to_insert, $blob_cols, $identity_col |
410 | ); |
411 | |
412 | $self->_identity($self->_writer_storage->_identity); |
413 | |
414 | $guard->commit; |
415 | |
416 | return $updated_cols; |
417 | } |
418 | |
419 | sub _insert { |
420 | my ($self, $next, $source, $to_insert, $blob_cols, $identity_col) = @_; |
421 | |
422 | my $updated_cols = $self->$next ($source, $to_insert); |
423 | |
424 | my $final_row = { |
425 | ($identity_col ? |
426 | ($identity_col => $self->last_insert_id($source, $identity_col)) : ()), |
427 | %$to_insert, |
428 | %$updated_cols, |
429 | }; |
430 | |
431 | $self->_insert_blobs ($source, $blob_cols, $final_row) if $blob_cols; |
432 | |
433 | return $updated_cols; |
434 | } |
435 | |
436 | sub update { |
437 | my $self = shift; |
438 | my ($source, $fields, $where, @rest) = @_; |
439 | |
440 | my $wantarray = wantarray; |
441 | |
442 | my $blob_cols = $self->_remove_blob_cols($source, $fields); |
443 | |
444 | my $table = $source->name; |
445 | |
6298a324 |
446 | my $identity_col = |
447 | first { $source->column_info($_)->{is_auto_increment} } $source->columns; |
057db5ce |
448 | |
449 | my $is_identity_update = $identity_col && defined $fields->{$identity_col}; |
450 | |
451 | return $self->next::method(@_) unless $blob_cols; |
452 | |
453 | # If there are any blobs in $where, Sybase will return a descriptive error |
454 | # message. |
455 | # XXX blobs can still be used with a LIKE query, and this should be handled. |
456 | |
457 | # update+blob update(s) done atomically on separate connection |
458 | $self = $self->_writer_storage; |
459 | |
460 | my $guard = $self->txn_scope_guard; |
461 | |
462 | # First update the blob columns to be updated to '' (taken from $fields, where |
463 | # it is originally put by _remove_blob_cols .) |
464 | my %blobs_to_empty = map { ($_ => delete $fields->{$_}) } keys %$blob_cols; |
465 | |
466 | # We can't only update NULL blobs, because blobs cannot be in the WHERE clause. |
467 | |
468 | $self->next::method($source, \%blobs_to_empty, $where, @rest); |
469 | |
470 | # Now update the blobs before the other columns in case the update of other |
471 | # columns makes the search condition invalid. |
472 | $self->_update_blobs($source, $blob_cols, $where); |
473 | |
474 | my @res; |
475 | if (%$fields) { |
476 | if ($wantarray) { |
477 | @res = $self->next::method(@_); |
478 | } |
479 | elsif (defined $wantarray) { |
480 | $res[0] = $self->next::method(@_); |
481 | } |
482 | else { |
483 | $self->next::method(@_); |
484 | } |
485 | } |
486 | |
487 | $guard->commit; |
488 | |
489 | return $wantarray ? @res : $res[0]; |
490 | } |
491 | |
492 | sub insert_bulk { |
493 | my $self = shift; |
494 | my ($source, $cols, $data) = @_; |
495 | |
6298a324 |
496 | my $identity_col = |
497 | first { $source->column_info($_)->{is_auto_increment} } $source->columns; |
057db5ce |
498 | |
6298a324 |
499 | my $is_identity_insert = (first { $_ eq $identity_col } @{$cols}) ? 1 : 0; |
057db5ce |
500 | |
501 | my @source_columns = $source->columns; |
502 | |
503 | my $use_bulk_api = |
504 | $self->_bulk_storage && |
505 | $self->_get_dbh->{syb_has_blk}; |
506 | |
507 | if ((not $use_bulk_api) |
508 | && |
509 | (ref($self->_dbi_connect_info->[0]) eq 'CODE') |
510 | && |
511 | (not $self->_bulk_disabled_due_to_coderef_connect_info_warned)) { |
512 | carp <<'EOF'; |
513 | Bulk API support disabled due to use of a CODEREF connect_info. Reverting to |
514 | regular array inserts. |
515 | EOF |
516 | $self->_bulk_disabled_due_to_coderef_connect_info_warned(1); |
517 | } |
518 | |
519 | if (not $use_bulk_api) { |
520 | my $blob_cols = $self->_remove_blob_cols_array($source, $cols, $data); |
521 | |
522 | # _execute_array uses a txn anyway, but it ends too early in case we need to |
523 | # select max(col) to get the identity for inserting blobs. |
524 | ($self, my $guard) = $self->{transaction_depth} == 0 ? |
525 | ($self->_writer_storage, $self->_writer_storage->txn_scope_guard) |
526 | : |
527 | ($self, undef); |
528 | |
529 | local $self->{insert_bulk} = 1; |
530 | |
531 | $self->next::method(@_); |
532 | |
533 | if ($blob_cols) { |
534 | if ($is_identity_insert) { |
535 | $self->_insert_blobs_array ($source, $blob_cols, $cols, $data); |
536 | } |
537 | else { |
538 | my @cols_with_identities = (@$cols, $identity_col); |
539 | |
540 | ## calculate identities |
541 | # XXX This assumes identities always increase by 1, which may or may not |
542 | # be true. |
543 | my ($last_identity) = |
544 | $self->_dbh->selectrow_array ( |
545 | $self->_fetch_identity_sql($source, $identity_col) |
546 | ); |
547 | my @identities = (($last_identity - @$data + 1) .. $last_identity); |
548 | |
549 | my @data_with_identities = map [@$_, shift @identities], @$data; |
550 | |
551 | $self->_insert_blobs_array ( |
552 | $source, $blob_cols, \@cols_with_identities, \@data_with_identities |
553 | ); |
554 | } |
555 | } |
556 | |
557 | $guard->commit if $guard; |
558 | |
559 | return; |
560 | } |
561 | |
562 | # otherwise, use the bulk API |
563 | |
564 | # rearrange @$data so that columns are in database order |
565 | my %orig_idx; |
566 | @orig_idx{@$cols} = 0..$#$cols; |
567 | |
568 | my %new_idx; |
569 | @new_idx{@source_columns} = 0..$#source_columns; |
570 | |
571 | my @new_data; |
572 | for my $datum (@$data) { |
573 | my $new_datum = []; |
574 | for my $col (@source_columns) { |
575 | # identity data will be 'undef' if not $is_identity_insert |
576 | # columns with defaults will also be 'undef' |
577 | $new_datum->[ $new_idx{$col} ] = |
578 | exists $orig_idx{$col} ? $datum->[ $orig_idx{$col} ] : undef; |
579 | } |
580 | push @new_data, $new_datum; |
581 | } |
582 | |
583 | # bcp identity index is 1-based |
584 | my $identity_idx = exists $new_idx{$identity_col} ? |
585 | $new_idx{$identity_col} + 1 : 0; |
586 | |
587 | ## Set a client-side conversion error handler, straight from DBD::Sybase docs. |
588 | # This ignores any data conversion errors detected by the client side libs, as |
589 | # they are usually harmless. |
590 | my $orig_cslib_cb = DBD::Sybase::set_cslib_cb( |
591 | Sub::Name::subname insert_bulk => sub { |
592 | my ($layer, $origin, $severity, $errno, $errmsg, $osmsg, $blkmsg) = @_; |
593 | |
594 | return 1 if $errno == 36; |
595 | |
596 | carp |
597 | "Layer: $layer, Origin: $origin, Severity: $severity, Error: $errno" . |
598 | ($errmsg ? "\n$errmsg" : '') . |
599 | ($osmsg ? "\n$osmsg" : '') . |
600 | ($blkmsg ? "\n$blkmsg" : ''); |
601 | |
602 | return 0; |
603 | }); |
604 | |
4edfce2f |
605 | my $exception = ''; |
ed7ab0f4 |
606 | try { |
057db5ce |
607 | my $bulk = $self->_bulk_storage; |
608 | |
609 | my $guard = $bulk->txn_scope_guard; |
610 | |
611 | ## XXX get this to work instead of our own $sth |
612 | ## will require SQLA or *Hacks changes for ordered columns |
613 | # $bulk->next::method($source, \@source_columns, \@new_data, { |
614 | # syb_bcp_attribs => { |
615 | # identity_flag => $is_identity_insert, |
616 | # identity_column => $identity_idx, |
617 | # } |
618 | # }); |
619 | my $sql = 'INSERT INTO ' . |
620 | $bulk->sql_maker->_quote($source->name) . ' (' . |
621 | # colname list is ignored for BCP, but does no harm |
622 | (join ', ', map $bulk->sql_maker->_quote($_), @source_columns) . ') '. |
623 | ' VALUES ('. (join ', ', ('?') x @source_columns) . ')'; |
624 | |
625 | ## XXX there's a bug in the DBD::Sybase bulk support that makes $sth->finish for |
626 | ## a prepare_cached statement ineffective. Replace with ->sth when fixed, or |
627 | ## better yet the version above. Should be fixed in DBD::Sybase . |
628 | my $sth = $bulk->_get_dbh->prepare($sql, |
629 | # 'insert', # op |
630 | { |
631 | syb_bcp_attribs => { |
632 | identity_flag => $is_identity_insert, |
633 | identity_column => $identity_idx, |
634 | } |
635 | } |
636 | ); |
637 | |
638 | my @bind = do { |
639 | my $idx = 0; |
640 | map [ $_, $idx++ ], @source_columns; |
641 | }; |
642 | |
643 | $self->_execute_array( |
644 | $source, $sth, \@bind, \@source_columns, \@new_data, sub { |
645 | $guard->commit |
646 | } |
647 | ); |
648 | |
649 | $bulk->_query_end($sql); |
ed7ab0f4 |
650 | } catch { |
651 | $exception = shift; |
057db5ce |
652 | }; |
653 | |
057db5ce |
654 | DBD::Sybase::set_cslib_cb($orig_cslib_cb); |
655 | |
656 | if ($exception =~ /-Y option/) { |
657 | carp <<"EOF"; |
658 | |
659 | Sybase bulk API operation failed due to character set incompatibility, reverting |
660 | to regular array inserts: |
661 | |
662 | *** Try unsetting the LANG environment variable. |
663 | |
664 | $exception |
665 | EOF |
666 | $self->_bulk_storage(undef); |
667 | unshift @_, $self; |
668 | goto \&insert_bulk; |
669 | } |
670 | elsif ($exception) { |
671 | # rollback makes the bulkLogin connection unusable |
672 | $self->_bulk_storage->disconnect; |
673 | $self->throw_exception($exception); |
674 | } |
675 | } |
676 | |
677 | sub _dbh_execute_array { |
678 | my ($self, $sth, $tuple_status, $cb) = @_; |
679 | |
680 | my $rv = $self->next::method($sth, $tuple_status); |
681 | $cb->() if $cb; |
682 | |
683 | return $rv; |
684 | } |
685 | |
686 | # Make sure blobs are not bound as placeholders, and return any non-empty ones |
687 | # as a hash. |
688 | sub _remove_blob_cols { |
689 | my ($self, $source, $fields) = @_; |
690 | |
691 | my %blob_cols; |
692 | |
693 | for my $col (keys %$fields) { |
694 | if ($self->_is_lob_column($source, $col)) { |
695 | my $blob_val = delete $fields->{$col}; |
696 | if (not defined $blob_val) { |
697 | $fields->{$col} = \'NULL'; |
698 | } |
699 | else { |
700 | $fields->{$col} = \"''"; |
701 | $blob_cols{$col} = $blob_val unless $blob_val eq ''; |
702 | } |
703 | } |
704 | } |
705 | |
706 | return %blob_cols ? \%blob_cols : undef; |
707 | } |
708 | |
709 | # same for insert_bulk |
710 | sub _remove_blob_cols_array { |
711 | my ($self, $source, $cols, $data) = @_; |
712 | |
713 | my @blob_cols; |
714 | |
715 | for my $i (0..$#$cols) { |
716 | my $col = $cols->[$i]; |
717 | |
718 | if ($self->_is_lob_column($source, $col)) { |
719 | for my $j (0..$#$data) { |
720 | my $blob_val = delete $data->[$j][$i]; |
721 | if (not defined $blob_val) { |
722 | $data->[$j][$i] = \'NULL'; |
723 | } |
724 | else { |
725 | $data->[$j][$i] = \"''"; |
726 | $blob_cols[$j][$i] = $blob_val |
727 | unless $blob_val eq ''; |
728 | } |
729 | } |
730 | } |
731 | } |
732 | |
733 | return @blob_cols ? \@blob_cols : undef; |
734 | } |
735 | |
736 | sub _update_blobs { |
737 | my ($self, $source, $blob_cols, $where) = @_; |
738 | |
9780718f |
739 | my @primary_cols = try |
740 | { $source->_pri_cols } |
741 | catch { |
742 | $self->throw_exception("Cannot update TEXT/IMAGE column(s): $_") |
743 | }; |
057db5ce |
744 | |
745 | # check if we're updating a single row by PK |
746 | my $pk_cols_in_where = 0; |
747 | for my $col (@primary_cols) { |
748 | $pk_cols_in_where++ if defined $where->{$col}; |
749 | } |
750 | my @rows; |
751 | |
752 | if ($pk_cols_in_where == @primary_cols) { |
753 | my %row_to_update; |
754 | @row_to_update{@primary_cols} = @{$where}{@primary_cols}; |
755 | @rows = \%row_to_update; |
756 | } else { |
757 | my $cursor = $self->select ($source, \@primary_cols, $where, {}); |
758 | @rows = map { |
759 | my %row; @row{@primary_cols} = @$_; \%row |
760 | } $cursor->all; |
761 | } |
762 | |
763 | for my $row (@rows) { |
764 | $self->_insert_blobs($source, $blob_cols, $row); |
765 | } |
766 | } |
767 | |
768 | sub _insert_blobs { |
769 | my ($self, $source, $blob_cols, $row) = @_; |
770 | my $dbh = $self->_get_dbh; |
771 | |
772 | my $table = $source->name; |
773 | |
774 | my %row = %$row; |
9780718f |
775 | my @primary_cols = try |
776 | { $source->_pri_cols } |
777 | catch { |
778 | $self->throw_exception("Cannot update TEXT/IMAGE column(s): $_") |
779 | }; |
057db5ce |
780 | |
781 | $self->throw_exception('Cannot update TEXT/IMAGE column(s) without primary key values') |
782 | if ((grep { defined $row{$_} } @primary_cols) != @primary_cols); |
783 | |
784 | for my $col (keys %$blob_cols) { |
785 | my $blob = $blob_cols->{$col}; |
786 | |
787 | my %where = map { ($_, $row{$_}) } @primary_cols; |
788 | |
789 | my $cursor = $self->select ($source, [$col], \%where, {}); |
790 | $cursor->next; |
791 | my $sth = $cursor->sth; |
792 | |
793 | if (not $sth) { |
057db5ce |
794 | $self->throw_exception( |
795 | "Could not find row in table '$table' for blob update:\n" |
6298a324 |
796 | . (Dumper \%where) |
057db5ce |
797 | ); |
798 | } |
799 | |
9780718f |
800 | try { |
057db5ce |
801 | do { |
802 | $sth->func('CS_GET', 1, 'ct_data_info') or die $sth->errstr; |
803 | } while $sth->fetch; |
804 | |
805 | $sth->func('ct_prepare_send') or die $sth->errstr; |
806 | |
807 | my $log_on_update = $self->_blob_log_on_update; |
808 | $log_on_update = 1 if not defined $log_on_update; |
809 | |
810 | $sth->func('CS_SET', 1, { |
811 | total_txtlen => length($blob), |
812 | log_on_update => $log_on_update |
813 | }, 'ct_data_info') or die $sth->errstr; |
814 | |
815 | $sth->func($blob, length($blob), 'ct_send_data') or die $sth->errstr; |
816 | |
817 | $sth->func('ct_finish_send') or die $sth->errstr; |
9780718f |
818 | } |
819 | catch { |
057db5ce |
820 | if ($self->using_freetds) { |
821 | $self->throw_exception ( |
9780718f |
822 | "TEXT/IMAGE operation failed, probably because you are using FreeTDS: $_" |
057db5ce |
823 | ); |
9780718f |
824 | } |
825 | else { |
826 | $self->throw_exception($_); |
057db5ce |
827 | } |
828 | } |
9780718f |
829 | finally { |
830 | $sth->finish if $sth; |
831 | }; |
057db5ce |
832 | } |
833 | } |
834 | |
835 | sub _insert_blobs_array { |
836 | my ($self, $source, $blob_cols, $cols, $data) = @_; |
837 | |
838 | for my $i (0..$#$data) { |
839 | my $datum = $data->[$i]; |
840 | |
841 | my %row; |
842 | @row{ @$cols } = @$datum; |
843 | |
844 | my %blob_vals; |
845 | for my $j (0..$#$cols) { |
846 | if (exists $blob_cols->[$i][$j]) { |
847 | $blob_vals{ $cols->[$j] } = $blob_cols->[$i][$j]; |
848 | } |
849 | } |
850 | |
851 | $self->_insert_blobs ($source, \%blob_vals, \%row); |
852 | } |
853 | } |
854 | |
855 | =head2 connect_call_datetime_setup |
856 | |
857 | Used as: |
858 | |
859 | on_connect_call => 'datetime_setup' |
860 | |
8384a713 |
861 | In L<connect_info|DBIx::Class::Storage::DBI/connect_info> to set: |
057db5ce |
862 | |
863 | $dbh->syb_date_fmt('ISO_strict'); # output fmt: 2004-08-21T14:36:48.080Z |
864 | $dbh->do('set dateformat mdy'); # input fmt: 08/13/1979 18:08:55.080 |
865 | |
866 | On connection for use with L<DBIx::Class::InflateColumn::DateTime>, using |
867 | L<DateTime::Format::Sybase>, which you will need to install. |
868 | |
869 | This works for both C<DATETIME> and C<SMALLDATETIME> columns, although |
870 | C<SMALLDATETIME> columns only have minute precision. |
871 | |
872 | =cut |
873 | |
874 | { |
875 | my $old_dbd_warned = 0; |
876 | |
877 | sub connect_call_datetime_setup { |
878 | my $self = shift; |
879 | my $dbh = $self->_get_dbh; |
880 | |
881 | if ($dbh->can('syb_date_fmt')) { |
882 | # amazingly, this works with FreeTDS |
883 | $dbh->syb_date_fmt('ISO_strict'); |
884 | } elsif (not $old_dbd_warned) { |
885 | carp "Your DBD::Sybase is too old to support ". |
886 | "DBIx::Class::InflateColumn::DateTime, please upgrade!"; |
887 | $old_dbd_warned = 1; |
888 | } |
889 | |
890 | $dbh->do('SET DATEFORMAT mdy'); |
891 | |
892 | 1; |
893 | } |
894 | } |
895 | |
896 | sub datetime_parser_type { "DateTime::Format::Sybase" } |
897 | |
898 | # ->begin_work and such have no effect with FreeTDS but we run them anyway to |
899 | # let the DBD keep any state it needs to. |
900 | # |
901 | # If they ever do start working, the extra statements will do no harm (because |
902 | # Sybase supports nested transactions.) |
903 | |
904 | sub _dbh_begin_work { |
905 | my $self = shift; |
906 | |
907 | # bulkLogin=1 connections are always in a transaction, and can only call BEGIN |
908 | # TRAN once. However, we need to make sure there's a $dbh. |
909 | return if $self->_is_bulk_storage && $self->_dbh && $self->_began_bulk_work; |
910 | |
911 | $self->next::method(@_); |
912 | |
913 | if ($self->using_freetds) { |
914 | $self->_get_dbh->do('BEGIN TRAN'); |
915 | } |
916 | |
917 | $self->_began_bulk_work(1) if $self->_is_bulk_storage; |
918 | } |
919 | |
920 | sub _dbh_commit { |
921 | my $self = shift; |
922 | if ($self->using_freetds) { |
923 | $self->_dbh->do('COMMIT'); |
924 | } |
925 | return $self->next::method(@_); |
926 | } |
927 | |
928 | sub _dbh_rollback { |
929 | my $self = shift; |
930 | if ($self->using_freetds) { |
931 | $self->_dbh->do('ROLLBACK'); |
932 | } |
933 | return $self->next::method(@_); |
934 | } |
935 | |
936 | # savepoint support using ASE syntax |
937 | |
938 | sub _svp_begin { |
939 | my ($self, $name) = @_; |
940 | |
941 | $self->_get_dbh->do("SAVE TRANSACTION $name"); |
942 | } |
943 | |
944 | # A new SAVE TRANSACTION with the same name releases the previous one. |
945 | sub _svp_release { 1 } |
946 | |
947 | sub _svp_rollback { |
948 | my ($self, $name) = @_; |
949 | |
950 | $self->_get_dbh->do("ROLLBACK TRANSACTION $name"); |
951 | } |
952 | |
953 | 1; |
954 | |
955 | =head1 Schema::Loader Support |
956 | |
290da7d6 |
957 | As of version C<0.05000>, L<DBIx::Class::Schema::Loader> should work well with |
958 | most (if not all) versions of Sybase ASE. |
057db5ce |
959 | |
960 | =head1 FreeTDS |
961 | |
962 | This driver supports L<DBD::Sybase> compiled against FreeTDS |
963 | (L<http://www.freetds.org/>) to the best of our ability, however it is |
964 | recommended that you recompile L<DBD::Sybase> against the Sybase Open Client |
965 | libraries. They are a part of the Sybase ASE distribution: |
966 | |
967 | The Open Client FAQ is here: |
968 | L<http://www.isug.com/Sybase_FAQ/ASE/section7.html>. |
969 | |
970 | Sybase ASE for Linux (which comes with the Open Client libraries) may be |
971 | downloaded here: L<http://response.sybase.com/forms/ASE_Linux_Download>. |
972 | |
973 | To see if you're using FreeTDS check C<< $schema->storage->using_freetds >>, or run: |
974 | |
975 | perl -MDBI -le 'my $dbh = DBI->connect($dsn, $user, $pass); print $dbh->{syb_oc_version}' |
976 | |
977 | Some versions of the libraries involved will not support placeholders, in which |
978 | case the storage will be reblessed to |
979 | L<DBIx::Class::Storage::DBI::Sybase::ASE::NoBindVars>. |
980 | |
981 | In some configurations, placeholders will work but will throw implicit type |
982 | conversion errors for anything that's not expecting a string. In such a case, |
983 | the C<auto_cast> option from L<DBIx::Class::Storage::DBI::AutoCast> is |
984 | automatically set, which you may enable on connection with |
985 | L<DBIx::Class::Storage::DBI::AutoCast/connect_call_set_auto_cast>. The type info |
986 | for the C<CAST>s is taken from the L<DBIx::Class::ResultSource/data_type> |
987 | definitions in your Result classes, and are mapped to a Sybase type (if it isn't |
988 | already) using a mapping based on L<SQL::Translator>. |
989 | |
48580715 |
990 | In other configurations, placeholders will work just as they do with the Sybase |
057db5ce |
991 | Open Client libraries. |
992 | |
993 | Inserts or updates of TEXT/IMAGE columns will B<NOT> work with FreeTDS. |
994 | |
995 | =head1 INSERTS WITH PLACEHOLDERS |
996 | |
997 | With placeholders enabled, inserts are done in a transaction so that there are |
998 | no concurrency issues with getting the inserted identity value using |
999 | C<SELECT MAX(col)>, which is the only way to get the C<IDENTITY> value in this |
1000 | mode. |
1001 | |
1002 | In addition, they are done on a separate connection so that it's possible to |
1003 | have active cursors when doing an insert. |
1004 | |
1005 | When using C<DBIx::Class::Storage::DBI::Sybase::ASE::NoBindVars> transactions |
1006 | are disabled, as there are no concurrency issues with C<SELECT @@IDENTITY> as |
1007 | it's a session variable. |
1008 | |
1009 | =head1 TRANSACTIONS |
1010 | |
48580715 |
1011 | Due to limitations of the TDS protocol, L<DBD::Sybase>, or both, you cannot |
1012 | begin a transaction while there are active cursors, nor can you use multiple |
057db5ce |
1013 | active cursors within a transaction. An active cursor is, for example, a |
1014 | L<ResultSet|DBIx::Class::ResultSet> that has been executed using C<next> or |
1015 | C<first> but has not been exhausted or L<reset|DBIx::Class::ResultSet/reset>. |
1016 | |
1017 | For example, this will not work: |
1018 | |
1019 | $schema->txn_do(sub { |
1020 | my $rs = $schema->resultset('Book'); |
1021 | while (my $row = $rs->next) { |
1022 | $schema->resultset('MetaData')->create({ |
1023 | book_id => $row->id, |
1024 | ... |
1025 | }); |
1026 | } |
1027 | }); |
1028 | |
1029 | This won't either: |
1030 | |
1031 | my $first_row = $large_rs->first; |
1032 | $schema->txn_do(sub { ... }); |
1033 | |
1034 | Transactions done for inserts in C<AutoCommit> mode when placeholders are in use |
1035 | are not affected, as they are done on an extra database handle. |
1036 | |
1037 | Some workarounds: |
1038 | |
1039 | =over 4 |
1040 | |
1041 | =item * use L<DBIx::Class::Storage::DBI::Replicated> |
1042 | |
1043 | =item * L<connect|DBIx::Class::Schema/connect> another L<Schema|DBIx::Class::Schema> |
1044 | |
1045 | =item * load the data from your cursor with L<DBIx::Class::ResultSet/all> |
1046 | |
1047 | =back |
1048 | |
1049 | =head1 MAXIMUM CONNECTIONS |
1050 | |
1051 | The TDS protocol makes separate connections to the server for active statements |
1052 | in the background. By default the number of such connections is limited to 25, |
1053 | on both the client side and the server side. |
1054 | |
1055 | This is a bit too low for a complex L<DBIx::Class> application, so on connection |
1056 | the client side setting is set to C<256> (see L<DBD::Sybase/maxConnect>.) You |
1057 | can override it to whatever setting you like in the DSN. |
1058 | |
1059 | See |
1060 | L<http://infocenter.sybase.com/help/index.jsp?topic=/com.sybase.help.ase_15.0.sag1/html/sag1/sag1272.htm> |
1061 | for information on changing the setting on the server side. |
1062 | |
1063 | =head1 DATES |
1064 | |
1065 | See L</connect_call_datetime_setup> to setup date formats |
1066 | for L<DBIx::Class::InflateColumn::DateTime>. |
1067 | |
1068 | =head1 TEXT/IMAGE COLUMNS |
1069 | |
1070 | L<DBD::Sybase> compiled with FreeTDS will B<NOT> allow you to insert or update |
1071 | C<TEXT/IMAGE> columns. |
1072 | |
1073 | Setting C<< $dbh->{LongReadLen} >> will also not work with FreeTDS use either: |
1074 | |
1075 | $schema->storage->dbh->do("SET TEXTSIZE $bytes"); |
1076 | |
1077 | or |
1078 | |
1079 | $schema->storage->set_textsize($bytes); |
1080 | |
1081 | instead. |
1082 | |
1083 | However, the C<LongReadLen> you pass in |
8384a713 |
1084 | L<connect_info|DBIx::Class::Storage::DBI/connect_info> is used to execute the |
1085 | equivalent C<SET TEXTSIZE> command on connection. |
057db5ce |
1086 | |
8384a713 |
1087 | See L</connect_call_blob_setup> for a |
1088 | L<connect_info|DBIx::Class::Storage::DBI/connect_info> setting you need to work |
1089 | with C<IMAGE> columns. |
057db5ce |
1090 | |
1091 | =head1 BULK API |
1092 | |
1093 | The experimental L<DBD::Sybase> Bulk API support is used for |
1094 | L<populate|DBIx::Class::ResultSet/populate> in B<void> context, in a transaction |
1095 | on a separate connection. |
1096 | |
1097 | To use this feature effectively, use a large number of rows for each |
1098 | L<populate|DBIx::Class::ResultSet/populate> call, eg.: |
1099 | |
1100 | while (my $rows = $data_source->get_100_rows()) { |
1101 | $rs->populate($rows); |
1102 | } |
1103 | |
1104 | B<NOTE:> the L<add_columns|DBIx::Class::ResultSource/add_columns> |
1105 | calls in your C<Result> classes B<must> list columns in database order for this |
1106 | to work. Also, you may have to unset the C<LANG> environment variable before |
1107 | loading your app, if it doesn't match the character set of your database. |
1108 | |
1109 | When inserting IMAGE columns using this method, you'll need to use |
1110 | L</connect_call_blob_setup> as well. |
1111 | |
6476fd66 |
1112 | =head1 COMPUTED COLUMNS |
1113 | |
1114 | If you have columns such as: |
1115 | |
1116 | created_dtm AS getdate() |
1117 | |
1118 | represent them in your Result classes as: |
1119 | |
1120 | created_dtm => { |
1121 | data_type => undef, |
1122 | default_value => \'getdate()', |
1123 | is_nullable => 0, |
1124 | } |
1125 | |
1126 | The C<data_type> must exist and must be C<undef>. Then empty inserts will work |
1127 | on tables with such columns. |
1128 | |
1129 | =head1 TIMESTAMP COLUMNS |
1130 | |
1131 | C<timestamp> columns in Sybase ASE are not really timestamps, see: |
1132 | L<http://dba.fyicenter.com/Interview-Questions/SYBASE/The_timestamp_datatype_in_Sybase_.html>. |
1133 | |
1134 | They should be defined in your Result classes as: |
1135 | |
1136 | ts => { |
1137 | data_type => 'timestamp', |
1138 | is_nullable => 0, |
1139 | inflate_datetime => 0, |
1140 | } |
1141 | |
1142 | The C<<inflate_datetime => 0>> is necessary if you use |
1143 | L<DBIx::Class::InflateColumn::DateTime>, and most people do, and still want to |
1144 | be able to read these values. |
1145 | |
1146 | The values will come back as hexadecimal. |
1147 | |
057db5ce |
1148 | =head1 TODO |
1149 | |
1150 | =over |
1151 | |
1152 | =item * |
1153 | |
1154 | Transitions to AutoCommit=0 (starting a transaction) mode by exhausting |
1155 | any active cursors, using eager cursors. |
1156 | |
1157 | =item * |
1158 | |
1159 | Real limits and limited counts using stored procedures deployed on startup. |
1160 | |
1161 | =item * |
1162 | |
1163 | Adaptive Server Anywhere (ASA) support, with possible SQLA::Limit support. |
1164 | |
1165 | =item * |
1166 | |
1167 | Blob update with a LIKE query on a blob, without invalidating the WHERE condition. |
1168 | |
1169 | =item * |
1170 | |
1171 | bulk_insert using prepare_cached (see comments.) |
1172 | |
1173 | =back |
1174 | |
1175 | =head1 AUTHOR |
1176 | |
1177 | See L<DBIx::Class/CONTRIBUTORS>. |
1178 | |
1179 | =head1 LICENSE |
1180 | |
1181 | You may distribute this code under the same terms as Perl itself. |
1182 | |
1183 | =cut |
1184 | # vim:sts=2 sw=2: |