BaseAdaptor.pm 19.6 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
=pod 

=head1 NAME

    Bio::EnsEMBL::Hive::DBSQL::BaseAdaptor

=head1 DESCRIPTION

    The base class for all other Object- or NakedTable- adaptors.
    Performs the low-level SQL needed to retrieve and store data in tables.

=head1 EXTERNAL DEPENDENCIES

    DBI 1.6

=head1 LICENSE

18
    Copyright [1999-2014] Wellcome Trust Sanger Institute and the EMBL-European Bioinformatics Institute
19
20
21
22
23
24
25
26
27
28
29
30

    Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License.
    You may obtain a copy of the License at

         http://www.apache.org/licenses/LICENSE-2.0

    Unless required by applicable law or agreed to in writing, software distributed under the License
    is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
    See the License for the specific language governing permissions and limitations under the License.

=head1 CONTACT

31
    Please subscribe to the Hive mailing list:  http://listserver.ebi.ac.uk/mailman/listinfo/ehive-users  to discuss Hive-related questions or to be notified of our updates
32
33
34
35

=cut


36
37
38
39
package Bio::EnsEMBL::Hive::DBSQL::BaseAdaptor;

use strict;
no strict 'refs';   # needed to allow AUTOLOAD create new methods
40
use DBI 1.6;        # the 1.6 functionality is important for detecting autoincrement fields and other magic.
41

42
43
use Bio::EnsEMBL::Hive::Utils ('stringify');

44
45
46
47
48
49
50
51
52
53
54

sub default_table_name {
    die "Please define table_name either by setting it via table_name() method or by redefining default_table_name() in your adaptor class";
}


sub default_insertion_method {
    return 'INSERT_IGNORE';
}


55
56
57
58
sub default_overflow_limit {
    return {
        # 'overflow_column1_name' => column1_size,
        # 'overflow_column2_name' => column2_size,
59
60
61
62
63
64
65
66
67
        # ...
    };
}

sub default_input_column_mapping {
    return {
        # 'original_column1' => "original_column1*10 AS c1_times_ten",
        # 'original_column2' => "original_column2+1 AS c2_plus_one",
        # ...
68
69
70
    };
}

71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
# ---------------------------------------------------------------------------

sub new {
    my ( $class, $dbobj ) = @_;

    my $self = bless {}, $class;

    if ( !defined $dbobj || !ref $dbobj ) {
        throw("Don't have a db [$dbobj] for new adaptor");
    }

    if ( ref($dbobj) =~ /DBConnection$/ ) {
        $self->dbc($dbobj);
    } elsif( UNIVERSAL::can($dbobj, 'dbc') ) {
        $self->dbc( $dbobj->dbc );
        $self->db( $dbobj );
    } else {
        throw("I was given [$dbobj] for a new adaptor");
    }

    return $self;
}


sub db {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_db} = shift @_;
    }
    return $self->{_db};
}


sub dbc {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_dbc} = shift @_;
    }
    return $self->{_dbc};
}


sub prepare {
    my ( $self, $sql ) = @_;

    # Uncomment next line to cancel caching on the SQL side.
    # Needed for timing comparisons etc.
    #$sql =~ s/SELECT/SELECT SQL_NO_CACHE/i;

    return $self->dbc->prepare($sql);
}

125
126
127
128
129
130
131
132
133
134
135

sub overflow_limit {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_overflow_limit} = shift @_;
    }
    return $self->{_overflow_limit} || $self->default_overflow_limit();
}


136
137
138
139
140
141
142
143
144
145
sub input_column_mapping {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_input_column_mapping} = shift @_;
    }
    return $self->{_input_column_mapping} || $self->default_input_column_mapping();
}


146
147
148
149
150
sub table_name {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_table_name} = shift @_;
151
        $self->_table_info_loader();
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
    }
    return $self->{_table_name} || $self->default_table_name();
}


sub insertion_method {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_insertion_method} = shift @_;
    }
    return $self->{_insertion_method} || $self->default_insertion_method();
}


sub column_set {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_column_set} = shift @_;
    } elsif( !defined( $self->{_column_set} ) ) {
        $self->_table_info_loader();
    }
    return $self->{_column_set};
}


sub primary_key {        # not necessarily auto-incrementing
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_primary_key} = shift @_;
    } elsif( !defined( $self->{_primary_key} ) ) {
        $self->_table_info_loader();
    }
    return $self->{_primary_key};
}


sub updatable_column_list {    # it's just a cashed view, you cannot set it directly
    my $self = shift @_;

    unless($self->{_updatable_column_list}) {
        my %primary_key_set = map { $_ => 1 } @{$self->primary_key()};
        my $column_set      = $self->column_set();
        $self->{_updatable_column_list} = [ grep { not $primary_key_set{$_} } keys %$column_set ];
    }
    return $self->{_updatable_column_list};
}


sub autoinc_id {
    my $self = shift @_;

    if(@_) {    # setter
        $self->{_autoinc_id} = shift @_;
    } elsif( !defined( $self->{_autoinc_id} ) ) {
        $self->_table_info_loader();
    }
    return $self->{_autoinc_id};
}


sub _table_info_loader {
    my $self = shift @_;

    my $dbc         = $self->dbc();
219
    my $dbh         = $dbc->db_handle();
220
    my $driver      = $dbc->driver();
221
    my $dbname      = $dbc->dbname();
222
223
224
225
    my $table_name  = $self->table_name();

    my %column_set  = ();
    my $autoinc_id  = '';
226
227
228
229
230
    my @primary_key = $dbh->primary_key(undef, undef, $table_name);

    my $sth = $dbh->column_info(undef, undef, $table_name, '%');
    $sth->execute();
    while (my $row = $sth->fetchrow_hashref()) {
231
232
233
        my ( $column_name, $column_type ) = @$row{'COLUMN_NAME', 'TYPE_NAME'};

        # warn "ColumnInfo [$table_name/$column_name] = $column_type\n";
234

235
        $column_set{$column_name}  = $column_type;
236

237
        if( ($column_name eq $table_name.'_id')
238
         or ($table_name eq 'analysis_base' and $column_name eq 'analysis_id') ) {    # a special case (historical)
239
            $autoinc_id = $column_name;
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
        }
    }
    $sth->finish;

    $self->column_set(  \%column_set );
    $self->primary_key( \@primary_key );
    $self->autoinc_id(   $autoinc_id );
}


sub count_all {
    my ($self, $constraint) = @_;

    my $table_name      = $self->table_name();

    my $sql = "SELECT COUNT(*) FROM $table_name";

257
258
259
    if($constraint) {
            # in case $constraint contains any kind of JOIN (regular, LEFT, RIGHT, etc) do not put WHERE in front:
        $sql .= (($constraint=~/\bJOIN\b/i) ? ' ' : ' WHERE ') . $constraint;
260
261
    }

262
    # warn "SQL: $sql\n";
263
264
265

    my $sth = $self->prepare($sql);
    $sth->execute;  
266
    my ($count) = $sth->fetchrow_array();
267
268
269
270
271
272
273
    $sth->finish;  

    return $count;
}


sub fetch_all {
274
275
    my ($self, $constraint, $one_per_key, $key_list, $value_column) = @_;
    
276
277
    my $table_name              = $self->table_name();
    my $input_column_mapping    = $self->input_column_mapping();
278

279
    my $sql = 'SELECT ' . join(', ', map { $input_column_mapping->{$_} // "$table_name.$_" } keys %{$self->column_set()}) . " FROM $table_name";
280
281

    if($constraint) { 
282
            # in case $constraint contains any kind of JOIN (regular, LEFT, RIGHT, etc) do not put WHERE in front:
283
        $sql .= (($constraint=~/\bJOIN\b/i or $constraint=~/^LIMIT|ORDER|GROUP/) ? ' ' : ' WHERE ') . $constraint;
284
285
    }

286
    # warn "SQL: $sql\n";
287
288
289
290

    my $sth = $self->prepare($sql);
    $sth->execute;  

291
292
293
    my @overflow_columns = keys %{ $self->overflow_limit() };
    my $overflow_adaptor = scalar(@overflow_columns) && $self->db->get_AnalysisDataAdaptor();

294
    my $result_struct;  # will be autovivified to the correct data structure
295
296

    while(my $hashref = $sth->fetchrow_hashref) {
297
298
299
300
301
302
303

        foreach my $overflow_key (@overflow_columns) {
            if($hashref->{$overflow_key} =~ /^_ext(?:\w+)_data_id (\d+)$/) {
                $hashref->{$overflow_key} = $overflow_adaptor->fetch_by_analysis_data_id_TO_data($1);
            }
        }

304
        my $pptr = \$result_struct;
305
306
307
308
        if($key_list) {
            foreach my $syll (@$key_list) {
                $pptr = \$$pptr->{$hashref->{$syll}};   # using pointer-to-pointer to enforce same-level vivification
            }
309
310
311
312
313
314
        }
        my $object = $value_column
            ? $hashref->{$value_column}
            : $self->objectify($hashref);
        if($one_per_key) {
            $$pptr = $object;
315
        } else {
316
            push @$$pptr, $object;
317
318
319
320
        }
    }
    $sth->finish;  

321
    unless(defined($result_struct)) {
322
        if($key_list and scalar(@$key_list)) {
323
324
325
326
327
328
329
            $result_struct = {};
        } elsif(!$one_per_key) {
            $result_struct = [];
        }
    }

    return $result_struct;  # either listref or hashref is returned, depending on the call parameters
330
331
332
333
}


sub primary_key_constraint {
334
335
    my $self        = shift @_;
    my $sliceref    = shift @_;
336
337
338
339

    my $primary_key  = $self->primary_key();  # Attention: the order of primary_key columns of your call should match the order in the table definition!

    if(@$primary_key) {
340
        return join (' AND ', map { $primary_key->[$_]."='".$sliceref->[$_]."'" } (0..scalar(@$primary_key)-1));
341
342
343
344
345
346
347
348
349
350
    } else {
        my $table_name = $self->table_name();
        die "Table '$table_name' doesn't have a primary_key";
    }
}


sub fetch_by_dbID {
    my $self = shift @_;    # the rest in @_ should be primary_key column values

351
    return $self->fetch_all( $self->primary_key_constraint( \@_ ), 1 );
352
353
354
}


355
356
357
sub remove_all {    # remove entries by a constraint
    my $self        = shift @_;
    my $constraint  = shift @_ || 1;
358

359
    my $table_name  = $self->table_name();
360

361
    my $sql = "DELETE FROM $table_name WHERE $constraint";
362
363
364
365
366
367
    my $sth = $self->prepare($sql);
    $sth->execute();
    $sth->finish();
}


368
369
370
371
372
373
374
375
376
377
sub remove {    # remove the object by primary_key
    my $self        = shift @_;
    my $object      = shift @_;

    my $primary_key_constraint  = $self->primary_key_constraint( $self->slicer($object, $self->primary_key()) );

    return $self->remove_all( $primary_key_constraint );
}


378
379
380
381
382
383
384
385
386
387
388
389
390
sub update {    # update (some or all) non_primary columns from the primary
    my $self    = shift @_;
    my $object  = shift @_;    # the rest in @_ should be the column names to be updated

    my $table_name              = $self->table_name();
    my $primary_key_constraint  = $self->primary_key_constraint( $self->slicer($object, $self->primary_key()) );
    my $columns_to_update       = scalar(@_) ? \@_ : $self->updatable_column_list();
    my $values_to_update        = $self->slicer( $object, $columns_to_update );

    unless(@$columns_to_update) {
        die "There are no dependent columns to update, as everything seems to belong to the primary key";
    }

391
    my $sql = "UPDATE $table_name SET ".join(', ', map { "$_=?" } @$columns_to_update)." WHERE $primary_key_constraint";
392
    # warn "SQL: $sql\n";
393
    my $sth = $self->prepare($sql);
394
    # warn "VALUES_TO_UPDATE: ".join(', ', map { "'$_'" } @$values_to_update)."\n";
395
396
    $sth->execute( @$values_to_update);

397
398
399
    $sth->finish();
}

400

401
sub store_or_update_one {
402
    my ($self, $object, $filter_columns) = @_;
403

404
    #use Data::Dumper;
405
    if(UNIVERSAL::can($object, 'adaptor') and $object->adaptor and $object->adaptor==$self) {  # looks like it has been previously stored
406
        if( @{ $self->primary_key() } and @{ $self->updatable_column_list() } ) {
407
            $self->update( $object );
408
409
410
            #warn "store_or_update_one: updated [".(UNIVERSAL::can($object, 'toString') ? $object->toString : Dumper($object))."]\n";
        } else {
            #warn "store_or_update_one: non-updatable [".(UNIVERSAL::can($object, 'toString') ? $object->toString : Dumper($object))."]\n";
411
        }
412
    } elsif( my $present = $self->check_object_present_in_db_by_content( $object, $filter_columns ) ) {
413
        $self->mark_stored($object, $present);
414
415
416
417
418
        #warn "store_or_update_one: found [".(UNIVERSAL::can($object, 'toString') ? $object->toString : Dumper($object))."] in db by content of (".join(', ', @$filter_columns).")\n";
        if( @{ $self->primary_key() } and @{ $self->updatable_column_list() } ) {
            #warn "store_or_update_one: updating the columns (".join(', ', @{ $self->updatable_column_list() }).")\n";
            $self->update( $object );
        }
419
420
    } else {
        $self->store( $object );
421
        #warn "store_or_update_one: stored [".(UNIVERSAL::can($object, 'toString') ? $object->toString : Dumper($object))."]\n";
422
423
424
    }
}

425

426
427
sub check_object_present_in_db_by_content {    # return autoinc_id/undef if the table has autoinc_id or just 1/undef if not
    my ( $self, $object, $filter_columns ) = @_;
428
429
430
431
432

    my $table_name  = $self->table_name();
    my $column_set  = $self->column_set();
    my $autoinc_id  = $self->autoinc_id();

433
434
435
436
437
438
439
440
441
    if($filter_columns) {
            # make sure all fields exist in the database as columns:
        $filter_columns = [ map { $column_set->{$_} ? $_ : $_.'_id' } @$filter_columns ];
    } else {
            # we look for identical contents, so must skip the autoinc_id columns when fetching:
        $filter_columns = [ grep { $_ ne $autoinc_id } keys %$column_set ];
    }
    my %filter_hash;
    @filter_hash{ @$filter_columns } = @{ $self->slicer( $object, $filter_columns ) };
442

443
444
    my @constraints = ();
    my @values = ();
445
    while(my ($column, $value) = each %filter_hash ) {
446
447
448
449
450
451
452
        if( defined($value) ) {
            push @constraints, "$column = ?";
            push @values, $value;
        } else {
            push @constraints, "$column IS NULL";
        }
    }
453

454
    my $sql = 'SELECT '.($autoinc_id or 1)." FROM $table_name WHERE ".  join(' AND ', @constraints);
455
#warn "check_object_present_in_db_by_content: sql= $sql WITH VALUES (".join(', ', @values).")\n";
456
457
    my $sth = $self->prepare( $sql );
    $sth->execute( @values );
458

459
    my ($return_value) = $sth->fetchrow_array();
460
461
462
463
464
465
466
    $sth->finish;

    return $return_value;
}


sub store {
467
    my ($self, $object_or_list) = @_;
468
469
470
471
472
473

    my $objects = (ref($object_or_list) eq 'ARRAY')     # ensure we get an array of objects to store
        ? $object_or_list
        : [ $object_or_list ];
    return unless(scalar(@$objects));

474
475
    my $table_name              = $self->table_name();
    my $autoinc_id              = $self->autoinc_id();
476
    my $all_storable_columns    = [ grep { $_ ne $autoinc_id } keys %{ $self->column_set() } ];
477
478
479
    my $driver                  = $self->dbc->driver();
    my $insertion_method        = $self->insertion_method;  # INSERT, INSERT_IGNORE or REPLACE
    $insertion_method           =~ s/_/ /g;
480
481
    if($driver eq 'sqlite') {
        $insertion_method =~ s/INSERT IGNORE/INSERT OR IGNORE/ig;
482
483
    } elsif($driver eq 'pgsql') {   # FIXME! temporary hack
        $insertion_method = 'INSERT';
484
    }
485

486
    my %hashed_sth = ();  # do not prepare statements until there is a real need
487

488
489
    my $stored_this_time        = 0;

490
    foreach my $object (@$objects) {
491
            my ($columns_being_stored, $column_key) = (ref($object) eq 'HASH') ? $self->keys_to_columns($object) : ($all_storable_columns, '*all*');
492
            my ($columns_being_stored, $column_key) = $self->keys_to_columns($object);
493
            # warn "COLUMN_KEY='$column_key'\n";
494
495

            my $this_sth;
496

497
498
499
500
                # only prepare (once!) if we get here:
            unless($this_sth = $hashed_sth{$column_key}) {
                    # By using question marks we can insert true NULLs by setting corresponding values to undefs:
                my $sql = "$insertion_method INTO $table_name (".join(', ', @$columns_being_stored).') VALUES ('.join(',', (('?') x scalar(@$columns_being_stored))).')';
501
                # warn "STORE: $sql\n";
502
503
504
                $this_sth = $hashed_sth{$column_key} = $self->prepare( $sql ) or die "Could not prepare statement: $sql";
            }

505
            # warn "STORED_COLUMNS: ".stringify($columns_being_stored)."\n";
506
            my $values_being_stored = $self->slicer( $object, $columns_being_stored );
507
            # warn "STORED_VALUES: ".stringify($values_being_stored)."\n";
508

509
            my $return_code = $this_sth->execute( @$values_being_stored )
510
                    # using $return_code in boolean context allows to skip the value '0E0' ('no rows affected') that Perl treats as zero but regards as true:
511
                or die "Could not store fields\n\t{$column_key}\nwith data:\n\t(".join(',', @$values_being_stored).')';
Leo Gordon's avatar
Leo Gordon committed
512
            if($return_code > 0) {     # <--- for the same reason we have to be explicitly numeric here
513
514
                my $liid = $autoinc_id && $self->dbc->db_handle->last_insert_id(undef, undef, $table_name, $autoinc_id);
                $self->mark_stored($object, $liid );
515
                ++$stored_this_time;
516
517
518
            }
    }

519
520
521
    foreach my $sth (values %hashed_sth) {
        $sth->finish();
    }
522

523
    return ($object_or_list, $stored_this_time);
524
525
526
527
528
529
530
531
}


sub DESTROY { }   # to simplify AUTOLOAD

sub AUTOLOAD {
    our $AUTOLOAD;

532
533
534
535
536
    if($AUTOLOAD =~ /::fetch(_all)?(?:_by_(\w+?))?(?:_HASHED_FROM_(\w+?))?(?:_TO_(\w+?))?$/) {
        my $all             = $1;
        my $filter_string   = $2;
        my $key_string      = $3;
        my $value_column    = $4;
537
538
539
540

        my ($self) = @_;
        my $column_set = $self->column_set();

541
        my $filter_components = $filter_string && [ split(/_AND_/i, $filter_string) ];
542
543
544
545
546
        foreach my $column_name ( @$filter_components ) {
            unless($column_set->{$column_name}) {
                die "unknown column '$column_name'";
            }
        }
547
        my $key_components = $key_string && [ split(/_AND_/i, $key_string) ];
548
        foreach my $column_name ( @$key_components ) {
549
550
551
552
            unless($column_set->{$column_name}) {
                die "unknown column '$column_name'";
            }
        }
553
554
555
        if($value_column && !$column_set->{$value_column}) {
            die "unknown column '$value_column'";
        }
556

557
#        warn "Setting up '$AUTOLOAD' method\n";
558
559
560
561
562
563
564
565
566
        *$AUTOLOAD = sub {
            my $self = shift @_;
            return $self->fetch_all(
                join(' AND ', map { "$filter_components->[$_]='$_[$_]'" } 0..scalar(@$filter_components)-1),
                !$all,
                $key_components,
                $value_column
            );
        };
567
        goto &$AUTOLOAD;    # restart the new method
568

569
    } elsif($AUTOLOAD =~ /::count_all_by_(\w+)$/) {
570
        my $filter_string = $1;
571
572
573
574

        my ($self) = @_;
        my $column_set = $self->column_set();

575
576
577
578
579
        my $filter_components = $filter_string && [ split(/_AND_/i, $filter_string) ];
        foreach my $column_name ( @$filter_components ) {
            unless($column_set->{$column_name}) {
                die "unknown column '$column_name'";
            }
580
        }
581

582
#        warn "Setting up '$AUTOLOAD' method\n";
583
584
585
586
587
588
589
590
        *$AUTOLOAD = sub {
            my $self = shift @_;
            return $self->count_all(
                join(' AND ', map { "$filter_components->[$_]='$_[$_]'" } 0..scalar(@$filter_components)-1),
            );
        };
        goto &$AUTOLOAD;    # restart the new method

591
592
593
594
595
596
597
    } elsif($AUTOLOAD =~ /::remove_all_by_(\w+)$/) {
        my $filter_name = $1;

        my ($self) = @_;
        my $column_set = $self->column_set();

        if($column_set->{$filter_name}) {
598
#            warn "Setting up '$AUTOLOAD' method\n";
599
600
601
602
603
            *$AUTOLOAD = sub { my ($self, $filter_value) = @_; return $self->remove_all("$filter_name='$filter_value'"); };
            goto &$AUTOLOAD;    # restart the new method
        } else {
            die "unknown column '$filter_name'";
        }
604
    } elsif($AUTOLOAD =~ /::update_(\w+)$/) {
605
        my @columns_to_update = split(/_AND_/i, $1);
606
#        warn "Setting up '$AUTOLOAD' method\n";
607
608
609
        *$AUTOLOAD = sub { my ($self, $object) = @_; return $self->update($object, @columns_to_update); };
        goto &$AUTOLOAD;    # restart the new method
    } else {
610
        warn "sub '$AUTOLOAD' not implemented";
611
612
613
614
615
    }
}

1;