SqlHelper.pm 26.7 KB
Newer Older
1 2 3 4 5 6
package Bio::EnsEMBL::Utils::SqlHelper;

=pod

=head1 LICENSE

7
  Copyright (c) 1999-2011 The European Bioinformatics Institute and
8 9 10 11 12 13 14 15 16 17
  Genome Research Limited.  All rights reserved.

  This software is distributed under a modified Apache license.
  For license details, please see

    http://www.ensembl.org/info/about/code_licence.html

=head1 CONTACT

  Please email comments or questions to the public Ensembl
18
  developers list at <dev@ensembl.org>.
19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 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

  Questions may also be sent to the Ensembl help desk at
  <helpdesk@ensembl.org>.
  
=cut

=pod

=head1 NAME

Bio::EnsEMBL::Utils::SqlHelper

=head1 VERSION

$Revision$

=head1 SYNOPSIS

	use Bio::EnsEMBL::Utils::SqlHelper;

	my $helper = Bio::EnsEMBL::Utils::SqlHelper->new(-DB_CONNECTION => $dbc);
	my $arr_ref = $helper->execute(
		-SQL => 'select name, age from tab where col =?',
		-CALLBACK => sub {
			my @row = @{shift @_};
			return {name=>$row[0], age=>$row[1]};
		},
		-PARAMS => ['A']
	);

	use Data::Dumper;
	print Dumper($arr_ref), "\n";
	#Prints out [name=>'name', age=>1] maybe ....
	
	#For transactional work; only works if your MySQL table 
	#engine/database supports transactional work (such as InnoDB)
	$helper->transaction( -CALLBACK => sub {
	  if($helper->execute_single_result(-SQL => 'select count(*) from tab')) {
	    return $helper->execute_update('delete from tab);
	  }
	  else {
	    return $helper->batch(-SQL => 'insert into tab (?,?)', -DATA => [
	     [1,2],
	     [1,3],
	     [1,4]
	    ]);
	  }
	});

=head1 DESCRIPTION

Easier database interaction

=head1 COMMITTER

$Author$

=head1 METHODS

See subrotuines.

=cut

use warnings;
use strict;

use Bio::EnsEMBL::Utils::Argument qw(rearrange);
use Bio::EnsEMBL::Utils::Scalar qw(assert_ref check_ref);
use Bio::EnsEMBL::Utils::Exception qw(throw warning);
use English qw( -no_match_vars ); #Used for $PROCESS_ID
use Scalar::Util qw(weaken); #Used to not hold a strong ref to DBConnection

=pod

=head2 new()

  Arg [DB_CONNECTION] : DBConnection instance to use
  Returntype          : Instance of helper
97 98
  Exceptions          : If the object given as a DBConnection is not one or it
                        was undefined
99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
  Status              : Stable

Creates a new instance of this object.

	my $dba = get_dba('mydb'); # New DBAdaptor from somewhere
	my $helper = Bio::EnsEMBL::Utils::SqlHelper->new(-DB_CONNECTION => $dba->dbc());
	$helper->execute_update(-SQL => 'update tab set flag=?', -PARAMS => [1]);

=cut

sub new {
	my ( $class, @args ) = @_;
	
	my ($db_connection) = rearrange([qw(db_connection)], @args);
	 
	my $self = bless( {}, ref($class) || $class );
115 116
	throw('-DB_CONNECTION construction parameter was undefined.') 
	 unless defined $db_connection;
117 118 119 120 121 122 123
	$self->db_connection($db_connection);
	
	return $self;
}

=pod

124 125
=head2 db_connection()

126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164
  Arg [1]     : DBConnection instance to use
  Description : Sets and retrieves the DBConnection 
  Returntype  : DBConnection if set; otherwise undef
  Exceptions  : If the object given as a DBConnection is not one or if an 
                attempt is made to set the value more than once
  Status      : Stable

=cut

sub db_connection {
  my ($self, $db_connection) = @_;
  if(defined $db_connection) {
    if(exists $self->{db_connection}) {
      throw('Cannot reset the DBConnection object; already defined ');
    }
    assert_ref($db_connection, 'Bio::EnsEMBL::DBSQL::DBConnection');
    $self->{db_connection} = $db_connection;
    weaken $self->{db_connection};
  }
  return $self->{db_connection};
}

# --------- SQL Methods

=pod

=head2 execute() - Execute a SQL statement with a custom row handler

  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for mapping a row to a data point; 
                        leave blank for a default mapping to a 2D array
  Arg [USE_HASHREFS]  : If set to true will cause HashRefs to be returned 
                        to the callback & not ArrayRefs
  Arg [PARAMS]        : The binding parameters to the SQL statement
  Returntype : 2D array containing the return of the callback
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable

	my $arr_ref = $helper->execute(
165 166
		-SQL 		=> 'select a,b,c from tab where col =?',
		-CALLBACK 	=> sub {
167 168 169
			my @row = @{shift @_};
			return {A=>$row[0], B=>$row[1], C=>$row[2]};
		},
170
		-PARAMS 	=> ['A']
171 172 173 174
	);
	
	#Or with hashrefs
	my $arr_ref = $helper->execute(
175 176 177
		-SQL 			=> 'select a,b,c from tab where col =?',
		-USE_HASHREFS 	=> 1,
		-CALLBACK 		=> sub {
178 179
			my $row = shift @_;
			return {A=>$row->{a}, B=>$row->{b}, C=>$row->{c}};
180 181
		},
		-PARAMS 		=> ['A']
182 183 184 185 186 187 188 189 190 191
	);

Uses a callback defined by the C<sub> decalaration. Here we specify how the 
calling code will deal with each row of a database's result set. The sub 
can return any type of Object/hash/data structure you require.

Should you not specify a callback then a basic one will be assigned to you
which will return a 2D array structure e.g.

  my $arr_ref = $helper->execute(
192
		-SQL 	=> 'select a,b,c from tab where col =?',
193 194 195 196 197 198 199 200 201 202 203 204
		-PARAMS => ['A']
	);
	
This is equivalent to DBI's c<selectall_arrayref()> subroutine.

As an extension to this method you can write a closure subroutine which
takes in two parameters. The first is the array/hash reference & the second is
the statement handle used to execute. 99% of the time you will not need
it but there are occasions where you do need it. An example of usage would be:

	my $conn = get_conn(); #From somwewhere
	my $arr_ref = $conn->execute(
205 206 207
		-SQL 			=> 'select a,b,c from tab where col =?',
		-USE_HASHREFS 	=> 1,
		-CALLBACK 		=> sub {
208 209 210
			my ($row, $sth) = @_;
			#Then do something with sth
			return {A=>$row->[0], B=>$row->[1], C=>$row->[2]};
211 212
		},
		-PARAMS 		=> ['A']
213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259
	);

Any arguments to bind to the incoming statement. This can be a set of scalars
or a 2D array if you need to specify any kind of types of sql objects i.e.

	use DBI qw(:sql_types);

	my $conn = get_conn();
	my $arr_ref = $conn->execute(
		-SQL => 'select a,b,c from tab where col =? and num_col=? and other=?',
		-USE_HASHREFS => 1,
		-CALLBACK => sub {
			my @row = @{shift @_};
			return {A=>$row[0], B=>$row[1], C=>$row[2]};
		},
		-PARAMS => ['1', SQL_VARCHAR], [2, SQL_INTEGER], 'hello'
	);

Here we import DBI's sql types into our package and then pass in multiple
anonymous array references as parameters. Each param is tested in the input
and if it is detected to be an ARRAY reference we dereference the array and
run DBI's bind_param method. In fact you can see each part of the incoming
paramaters array as the contents to call C<bind_param> with. The only difference
is the package tracks the bind position for you.

=cut

sub execute {
	my ( $self, @args ) = @_;
	my ($sql, $callback, $use_hashrefs, $params) = rearrange([qw(sql callback use_hashrefs params)], @args);
	my $has_return = 1;
	
	#If no callback then we execute using a default one which returns a 2D array
	if(!defined $callback) {
    throw('Cannot use fetchrow_hashref() with default mappers. Turn off this option') if $use_hashrefs;
    $callback = $self->_mappers()->{array_ref};
	}
	
	return $self->_execute( $sql, $callback, $has_return, $use_hashrefs, $params );
}

=pod

=head2 execute_simple()

  Arg [SQL]           : SQL to execute
  Arg [PARAMS]        : The binding parameters to the SQL statement
260
  Arg [CALLBACK]      : Allows you to give a callback to do the mapping with
261 262 263 264 265
  Returntype : 1D array of data points
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable

  my $classification = $helper->execute_simple(
266
    -SQL 	=> 'select meta_val from meta where meta_key =? order by meta_id', 
267 268 269 270 271 272 273 274 275 276 277
    -PARAMS => ['species.classification']
  );

Identical to C<execute> except you do not specify a sub-routine reference. 
Using this code assumes you want an array of single scalar values as returned 
by the given SQL statement.

=cut

sub execute_simple {
  my ( $self, @args ) = @_;
278
	my ($sql, $params, $callback) = rearrange([qw(sql params callback)], @args);
279 280
	my $has_return = 1;
	my $use_hashrefs = 0;
281
	$callback ||= $self->_mappers()->{first_element};
282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323
	return $self->_execute($sql, $callback, $has_return, $use_hashrefs, $params);
}

=pod

=head2 execute_no_return()

  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for mapping a row to a data point;
                        we assume you are assigning into a data structure which
                        has requirements other than simple translation into an
                        array
  Arg [USE_HASHREFS]  : If set to true will cause HashRefs to be returned 
                        to the callback & not ArrayRefs
  Arg [PARAMS]        : The binding parameters to the SQL statement
  Returntype : None
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable

Whilst all other execute methods will return something; this assumes that the
given mapper subroutine will be performing the business of placing values
somewhere or doing something with them.

There is a huge temptation to nest queries using this method; do not! Execute
the values into an array using one of the other methods then run your subqueries
on them; or make a better first query. SQL is flexible; so use it.

=cut

sub execute_no_return {
	my ( $self, @args ) = @_;
	my ($sql, $callback, $use_hashrefs, $params) = rearrange([qw(sql callback use_hashrefs params)], @args);
	throw('No callback defined but this is a required parameter for execute_no_return()') if ! $callback;
	my $has_return = 0;
	$self->_execute( $sql, $callback, $has_return, $use_hashrefs, $params );
	return;
}

=pod

=head2 execute_into_hash()

324 325 326 327 328 329 330 331 332
  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for mapping to a value in a hash
                        keyed by the first element in your result set; 
                        leave blank for a default mapping to a scalar value
                        of the second element
  Arg [PARAMS]        : The binding parameters to the SQL statement
  Returntype : A HashRef keyed by column 1 & value is the return of callback
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable
333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400

A variant of the execute methods but rather than returning a list of mapped
results this will assume the first column of a returning map & the calling
subroutine will map the remainder of your return as the hash's key. For example:

	my $sql = 'select key, one, two from table where something =?';
	my $mapper = sub {
		my ($row) = @_;
		#Ignore field 0 as that is being used for the key
		my $obj = Some::Obj->new(one=>$row->[1], two=>$row->[2]);
		return $obj;
	};

	my $hash = $helper->execute_into_hash(-SQL => $sql, -CALLBACK => $mapper, -PARAMS => ['val']);
	
	#Or for a more simple usage
	my $sql = 'select biotype, count(gene_id) from gene group by biotype';
	my $biotype_hash = $conn->execute_into_hash(-SQL => $sql);
	print $biotype_hash->{protein_coding} || 0, "\n";

The basic pattern assumes a scenario where you are mapping in a one key to
one value. For more advanced mapping techniques you need to start using the
non-consuming executes which allow you to process a result set without assuming
that you want to map the rows into single objects.

B<Remember that the row you are given is the full row & not a view of the
reminaing fields.> Therefore indexing for the data you are concerned with
begins at position 1.

=cut

sub execute_into_hash {
	my ( $self, @args ) = @_;
	my ($sql, $callback, $params) = rearrange([qw(sql callback params)], @args);
	my $hash = {};
	
	#If no callback then we execute using a default one which sets value to 2nd element
	if(!defined $callback) {
	 $callback = $self->_mappers()->{second_element};
	}
	
	#Default mapper uses the 1st key + something else from the mapper
	my $mapper = sub {
		my $row   = shift @_;
		my $value = $callback->($row);
		$hash->{ $row->[0] } = $value;
		return;
	};
	
	$self->execute_no_return(
	  -SQL => $sql, 
	  -CALLBACK => $mapper,
	  -PARAMS => $params
	);
	
	return $hash;
}

=pod

=head2 execute_single_result()

  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for mapping a row to a data point; 
                        leave blank for a default scalar mapping
  Arg [USE_HASHREFS]  : If set to true will cause HashRefs to be returned 
                        to the callback & not ArrayRefs
  Arg [PARAMS]        : The binding parameters to the SQL statement
401 402 403
  Returntype : One data point
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442

  my $meta_count = $helper->execute_single_result(
    -SQL => 'select count(*) from meta where species_id =?', 
    -PARAMS => [1]
  );

Very similar to C<execute()> except it will raise an exception if we have more 
or less than one row returned

=cut

sub execute_single_result {
	my ( $self, @args ) = @_;
	my ($sql, $callback, $use_hashrefs, $params) = rearrange(
	 [qw(sql callback use_hashrefs params)], @args);
	
	my $results = $self->execute_simple( 
	  -SQL => $sql, 
	  -CALLBACK => $callback, 
	  -USE_HASHREFS => $use_hashrefs, 
	  -PARAMS => $params
	);
	
	my $result_count = scalar(@{$results});
	if($result_count != 1) {
	  $params = [] if ! $params;
	  my $type = ($result_count == 0) ? 'No' : 'Too many';
		my $msg = "${type} results returned. Expected 1 but got $result_count for query '${sql}' with params [";
		$msg .= join( ',', map {(defined $_) ? $_ : '-undef-';} @{$params} );
		$msg .= ']';
		throw($msg);
	}
	return $results->[0];
}

=pod

=head2 transaction()

443 444 445 446 447 448
  Arg [CALLBACK]      : The callback used for transaction isolation; once 
                        the subroutine exists the code will decide on rollback
                        or commit
  Returntype : Return of the callback
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable
449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471

  my $val = $helper->transaction(-CALLBACK => sub {
    my ($dbc) = @_;
    #Do something
    return 1;
  });
  
  #Or because of the arguments method we use
  my $val = $helper->transaction(sub {
    my ($dbc) = @_;
    #Do something
    return 1;
  });

Creates a transactional block which will ensure that the connection is committed
when your submmited subroutine has finished or will rollback in the event of
an error occuring in your block.

The code will always force AutoCommit off but will restore it to its 
previous setting. If your DBI/DBD driver does not support manual commits
then this code will break. The code will turn off the 
C<disconnect_when_idle()> method to allow transactions to work as expected.

472 473
An effect of using REPEATABLE READ transaction isolation (InnoDB's default) 
is that your data is as fresh as when you started your current transaction. To
474 475 476 477
ensure the freshest data use C<SELECT ... from ... LOCK IN SHARE MODE>
or C<SELECT ... from ... LOCK FOR UPDATE> if you are going to issue
updates.

478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
Creating a transaction within a transaction results in the commit rollback 
statements occuring in the top level transaction. That way any block of 
code which is meant to to be transaction can be wrapped in this block (
assuming the same instance of SQLHelper is passed around & used).

=cut

sub transaction {
  my ($self, @args) = @_;
  
  my ($callback) = rearrange([qw(callback)], @args);
  
  throw('Callback was not a CodeRef. Got a reference of type ['.ref($callback).']') 
    unless check_ref($callback, 'CODE');
 
  my $dbc = $self->db_connection();
  my $original_dwi;
  my $ac;
  
497 498 499
  my $error;
  my $result;
  
500 501 502 503 504
  #If we were already in a transaction then we do not do any management of the
  #session & wait for the parent transaction(s) to finish
  my $perform_transaction = $self->_perform_transaction_code();
  if($perform_transaction) {
    $original_dwi = $dbc->disconnect_when_inactive();
505
    $dbc->disconnect_when_inactive(0);
506 507
    $ac = $dbc->db_handle()->{'AutoCommit'};
    $dbc->db_handle()->{'AutoCommit'} = 0;
508
    $self->_enable_transaction();
509 510
  }
  
511 512 513 514 515 516 517
  if(!$error) {
    eval {
      $result = $callback->($dbc);
      $dbc->db_handle()->commit() if $perform_transaction;
    };
    $error = $@;
  }
518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
  
  if($perform_transaction) {
    if($error) {
      eval { $dbc->db_handle()->rollback(); };
    }
    $dbc->db_handle()->{'AutoCommit'} = $ac;
    $dbc->disconnect_when_inactive($original_dwi);
    $self->_disable_transaction();
  }
  
  throw("Transaction aborted because of error: ${error}") if $error;
  
  return $result;
}

=pod

=head2 execute_update()

537 538 539 540 541 542 543 544
  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for calling methods on the 
                        DBI statement handle or DBConnection object after an 
                        update command
  Arg [PARAMS]        : The binding parameters to the SQL statement
  Returntype : Number of rows affected
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable
545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562

Used for performing updates but conforms to the normal execute statement
subroutines.

	use DBI qw(:sql_types);
	$helper->execute_update(
	 -SQL => 'update tab set name = ? where id =?', 
	 -PARAMS => ['andy', [1, SQL_INTEGER]]
	);

If you need to do something a bit more advanced with your DML then you can
give the method a closure and this will be called after the execute has been
issued i.e.

	my $obj;
	$helper->execute_update(
	 -SQL => 'insert into tab (name) values(?)', 
	 -CALLBACK => sub {
563
	   my ($sth, $dbh) = @_;
564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
	   $obj->{id} = $dbh->{mysql_insertid);
   }, 
   -PARAMS => [$obj->name()]
  );

This lets us access the statement handle & database handle to access other
properties such as the last identifier inserted.

=cut

sub execute_update {
  my ($self, @args) = @_;
  my ($sql, $callback, $params) = rearrange([qw(sql callback params)], @args);
  my $rv = 0;
  my $sth;
  eval {
    $sth = $self->db_connection()->prepare($sql);
    $self->_bind_params($sth, $params);
    $rv = $sth->execute();
583
    $callback->($sth, $self->db_connection()->db_handle()) if $callback;
584 585 586 587 588 589 590 591 592 593
  };
  my $error = $@;
  $self->_finish_sth($sth);
  if($error) {
    my $params = join ' ', map { (defined $_) ? $_ : q{undef} } @{$params};
    throw("Cannot apply sql '${sql}' with params '${params}': ${error}");
  }
  return $rv;
}

594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638
=head2 execute_with_sth()

  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for working with the statement
                        handle once returned. This is B<not> a mapper.
  Arg [PARAMS]        : The binding parameters to the SQL statement
  Description : A subrotuine which abstracts resource handling and statement
                preparing leaving the developer to define how to handle
                and process the statement.
  Returntype  : Anything you wish to return from the callback
  Exceptions  : If errors occur in the execution of the SQL
  Status      : Stable

  my $meta_count = $helper->execute_with_sth(
    -SQL => 'select count(*) from meta where species_id =?', 
    -PARAMS => [1],
    -CALLBACK => sub {
      my ($sth) = @_;
      my $count;
      $sth->bind_columns(\($count));
      while ($sth->fetch) {
        print $count, "\n";
      }
      return $count;
    }
  );

Very similar to C<execute()> except this gives you full control over the
lifecycle of the statement handle & how you wish to proceed with working
with a statement handle. This is for situations where you believe going through
the mappers causes too much of a slow-down (since we have to execute a
subroutine for every row in order to map it correctly).

However please benchmark before adopting this method as it increases the 
complexity of your code and the mapper slow down only becomes apparent when
working with very large numbers of rows.

=cut

sub execute_with_sth {
  my ($self, @args) = @_;
  my ($sql, $callback, $params) = rearrange([qw(sql callback params)], @args);
  return $self->_base_execute( $sql, 1, $params, $callback );
}

639 640
=pod

Andy Yates's avatar
Bad POD  
Andy Yates committed
641
=head2 batch()
642

643 644 645 646 647 648 649 650 651 652 653 654 655
  Arg [SQL]           : SQL to execute
  Arg [CALLBACK]      : The callback to use for working with the statement
                        handle once returned; specify this or -DATA
  Arg [DATA]          : The data to insert; specify this or -CALLBACK
  Arg [COMMIT_EVERY]  : Integer; defines the rate at which to issue commits to
                        the DB handle. This is important when working with 
                        InnoDB databases since it affects the speed of rollback
                        (larger gaps inbetween commits means more to rollback).
                        
                        Ignored if using the callback version.
  Returntype : Numbers of rows updated
  Exceptions : If errors occur in the execution of the SQL
  Status     : Stable
656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690

  my $alotofdata = getitfromsomewhere();
  $helper->batch(-SQL => 'insert into table (one,two) values(?,?)', -CALLBACk => sub {
    my ($sth, $dbc) = @_;
    foreach my $data (@alotofdata) {
      $sth->execute(@{$data});
    }
  });
  
  #Or for a 2D array data driven approach
  $helper->batch(-SQL => 'insert into table (one,two) values(?,?)', -DATA => $alotofdata);

Takes in a sql statement & a code reference. Your SQL is converted into a 
prepared statement & then given as the first parameter to the closure. The
second parameter is the DBH which created the statement. This is intended
to let you do mass insertion into a database without the need to
re-preparing the same statement.

This can be combined with the transaction() code to provide a construct
which does batch insertion & is transactionally aware.

We can also use data based batch insertions i.e.

  #Needs to be like:
  #   [ [1,2], [3,4] ]
  #Or if using the DBI types: 
  #   [ [ [1, SQL_INTEGER], [2, SQL_INTEGER] ], [ [3, SQL_INTEGER], [4, SQL_INTEGER] ] ]
  my $alotofdata = getitfromsomewhere(); 
  $helper->batch(-SQL => 'insert into table (one,two) values(?,?)', 
    -DATA => $alotofdata);
  
This does exactly what the previous example.

All batch statements will return the value the callback computes. If you are 
using the previous example with a data array then the code will return the
691
number affected rows by the query.
692 693 694 695 696

=cut

sub batch {
  my ($self, @args) = @_;
697
  my ($sql, $callback, $data, $commit_every) = rearrange([qw(sql callback data commit_every)], @args);
698 699 700 701 702 703 704 705 706 707
  
  if(! defined $callback && ! defined $data) {
    throw('You need to define a callback for insertion work or the 2D data array');
  }
  
  my $result;
  if(defined $callback) {
    $result = $self->_callback_batch($sql, $callback);
  }
  else {
708
    $result = $self->_data_batch($sql, $data, $commit_every);
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777
  }
  return $result if defined $result;
  return;
}

#------- Internal methods

sub _mappers {
  my ($self) = @_;
  if(! exists $self->{_mappers}) {
    $self->{_mappers} = {
      first_element => sub {
        my ($row) = @_;
        return $row->[0];
      },
      second_element => sub {
        my ($row) = @_;
        return $row->[1];
      },
      array_ref => sub {
        my $row = shift @_;
	     return [@{$row}];
      }
    };
  }
  return $self->{_mappers};
}

sub _perform_transaction_code {
  my ($self) = @_;
  return $self->{_transaction_active}->{$PROCESS_ID} ? 0 : 1;
}

sub _enable_transaction {
  my ($self) = @_;
  $self->{_transaction_active}->{$PROCESS_ID} = 1;
  return;
}

sub _disable_transaction {
  my ($self) = @_;
  delete $self->{_transaction_active}->{$PROCESS_ID};
  return;
}

sub _bind_params {
	my ( $self, $sth, $params ) = @_;
	
	return if ! defined $params; #Return quickly if we had no data
	
	if(! check_ref($params, 'ARRAY')) {
	  throw(qq{The given parameters reference '${params}' is not an ARRAY; wrap in an ArrayRef});
	}
	
	my $count = 1;
	foreach my $param (@{$params}) {
		if ( check_ref($param, 'ARRAY') ) {
			$sth->bind_param( $count, @{$param} );
		}
		else {
			$sth->bind_param( $count, $param );
		}
		$count++;
	}
	return;
}

sub _execute {
	my ( $self, $sql, $callback, $has_return, $use_hashrefs, $params ) = @_;
778

779 780
	throw('Not given a mapper. _execute() must always been given a CodeRef') unless check_ref($callback, 'CODE');
	
781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813
  my @results;
  
  my $sth_processor;
  if($use_hashrefs) {
    
    $sth_processor = sub {
      my ($sth) = @_;
      while( my $row = $sth->fetchrow_hashref() ) {
		    push(@results, $callback->($row, $sth));
		  }
    };
  }
  else {
    
    $sth_processor = sub {
      my ($sth) = @_;
      while ( my $row = $sth->fetchrow_arrayref() ) {
  			push(@results, $callback->($row, $sth));
  		}
    };
  }
  
  $self->_base_execute($sql, $has_return, $params, $sth_processor);
  
  return \@results if $has_return;
	return;
}

sub _base_execute {
  my ( $self, $sql, $has_return, $params, $sth_processor ) = @_;
  
  throw('Not given a sth_processor. _base_execute() must always been given a CodeRef') unless check_ref($sth_processor, 'CODE');
	
814 815 816 817 818 819 820 821 822 823 824 825 826 827
	$params = [] unless $params;
	
	my $conn = $self->db_connection;
	my @results;

	my $error;
	my $sth_close_error;
	my $sth;

	eval {
		$sth = $conn->prepare($sql);
		throw("Cannot continue as prepare() did not return a handle") unless $sth;
		$self->_bind_params( $sth, $params );
		$sth->execute();
828
    	$sth_processor->($sth);
829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864
	};
	
	$error = $@;
	$self->_finish_sth($sth);
	if($error) {
  	throw("Cannot run '${sql}' with params '@{$params}' due to error: $error") if $error;
	}
	return \@results if $has_return;
	return;
}

sub _finish_sth {
  my ($self, $sth) = @_;
  eval { $sth->finish() if defined $sth; };
  warning('Cannot finish() the statement handle: $@') if $@;
  return;
}

sub _callback_batch {
  my ($self, $sql, $callback) = @_;
  my $error;
  my $sth;
  my $closure_return;
  eval {
    $sth = $self->db_connection()->prepare($sql); 
    $closure_return = $callback->($sth, $self->db_connection());
  };
  $error = $@;
  $self->_finish_sth($sth);
	throw("Problem detected during batch work: $error") if $error;
  
  return $closure_return if defined $closure_return;
  return;
}

sub _data_batch {
865
  my ($self, $sql, $data, $commit_every) = @_;
866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886
  
  #Input checks
  assert_ref($data, 'ARRAY');
  my $data_length = scalar(@{$data});
  return 0 unless $data_length > 0;
  my $first_row = $data->[0];
  throw('I expect to work with a 2D ArrayRef but this is not one') unless check_ref($first_row, 'ARRAY');

  my $callback = sub {
    my ($sth, $dbc) = @_;
    my $total_affected = 0;
    #Iterate over each data point
    for(my $data_index = 0; $data_index < $data_length; $data_index++) {
      my $row = $data->[$data_index];
      $self->_bind_params($sth, $row);
      my $affected = eval {$sth->execute()};
      if($@) {
        throw("Problem working with $sql with params @{$row}: $@");
      }
      my $num_affected = ($affected) ? $affected :  0; #Get around DBI's 0E0
      $total_affected += $num_affected;
887 888 889 890 891 892 893 894 895 896 897 898 899
      
      #Lets us do a commit once every x rows apart from 0. We also finish
      #off with a commit if the code told us we were doing it
      if($commit_every) {
        if( ($data_index % $commit_every == 0) && $data_index != 0) {
          $dbc->db_handle()->commit();
        }
      }
    }
    
    #finish off with a commit if the code told us we were doing it
    if($commit_every) {
      $dbc->db_handle()->commit();
900
    }
901
    
902 903 904 905 906 907 908
    return $total_affected || 0;
  };
  
  return $self->_callback_batch($sql, $callback)
}

1;