LICENSE
Copyright [1999-2015] Wellcome Trust Sanger Institute and the EMBL-European Bioinformatics Institute Copyright [2016-2024] EMBL-European Bioinformatics Institute
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-2.0Unless 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.
CONTACT
Please email comments or questions to the public Ensembl
developers list at <http://lists.ensembl.org/mailman/listinfo/dev>.
Questions may also be sent to the Ensembl help desk at
<http://www.ensembl.org/Help/Contact>.NAME
Bio::EnsEMBL::Utils::SqlHelper
VERSION
$Revision$
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 ] ] );
}
} );DESCRIPTION
Easier database interaction
new()
Arg [DB_CONNECTION] : Bio::EnsEMBL::DBSQL::DBConnection $db_connection
Returntype : Instance of helper
Exceptions : If the object given as a DBConnection is not one or it
was undefined
Status : Stable
Description : Creates a new instance of this object.
Example :
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] );db_connection()
Arg [1] : Bio::EnsEMBL::DBSQL::DBConnection $db_connection
Description : Sets and retrieves the DBConnection
Returntype : Bio::EnsEMBL::DBSQL::DBConnection
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 : Stableexecute() - Execute a SQL statement with a custom row handler
Arg [SQL] : string SQL to execute
Arg [CALLBACK] : CodeRef; 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] : boolean If set to true will cause HashRefs to be returned
to the callback & not ArrayRefs
Arg [PARAMS] : ArrayRef The binding parameters to the SQL statement
Arg [PREPARE_PARAMS] : boolean Parameters to be passed onto the Statement Handle
prepare call
Arg [ITERATOR] : boolean Request a Bio::EnsEMBL::Utils::Iterator
rather than a 2D array
Returntype : ArrayRef or Bio::EnsEMBL::Utils::Iterator
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Example :
my $arr_ref = $helper->execute(
-SQL => 'select a,b,c from tab where col =?',
-CALLBACK => sub {
my @row = @{ shift @_ };
return { A => $row[0], B => $row[1], C => $row[2] };
},
-PARAMS => ['A'] );
#Or with hashrefs
my $arr_ref = $helper->execute(
-SQL => 'select a,b,c from tab where col =?',
-USE_HASHREFS => 1,
-CALLBACK => sub {
my $row = shift @_;
return { A => $row->{a}, B => $row->{b}, C => $row->{c} };
},
-PARAMS => ['A'] );
Description:
Uses a callback defined by the 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(
-SQL => 'select a,b,c from tab where col =?',
-PARAMS => ['A'] );This is equivalent to DBI's 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(
-SQL => 'select a,b,c from tab where col =?',
-USE_HASHREFS => 1,
-CALLBACK => sub {
my ( $row, $sth ) = @_;
#Then do something with sth
return { A => $row->[0], B => $row->[1], C => $row->[2] };
},
-PARAMS => ['A'] );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 bind_param with. The only difference is the package tracks the bind position for you.
We can get back a Bio::EnsEMBL::Utils::Iterator object which can be used to iterate over the results set without first materializing the data into memory. An example would be:
my $iterator = $helper->execute(
-SQL => 'select a,b,c from tab where col =?',
-PARAMS => ['A']
-ITERATOR => 1);
while($iterator->has_next()) {
my $row = $iterator->next();
#Do something
}This is very useful for very large datasets.
execute_simple()
Arg [SQL] : string $sql
Arg [PARAMS] : ArrayRef $params
Arg [CALLBACK] : CodeRef $callback
Returntype : ArrayRef of 1D elements
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Example : my $classification =
$helper->execute_simple(
-SQL => 'select meta_val from meta where meta_key =? order by meta_id',
-PARAMS => ['species.classification']
);
Description: Similar to execute() but without a sub-routine reference.
Using this code assumes you want an array of single scalar values
as returned by the given SQL statement.execute_no_return()
Arg [SQL] : string sql
Arg [CALLBACK] : CodeRef 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] : boolean If set to true will cause HashRefs to be returned
to the callback and not ArrayRefs
Arg [PARAMS] : ArrayRef The binding parameters to the SQL statement
Returntype : None
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Description: 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.execute_into_hash()
Arg [SQL] : string $sql
Arg [CALLBACK] : CodeRef 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 : HashRef keyed by column 1 & value is the return of callback
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Description: 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 and
the calling subroutine will map the remainder of your return as the
hash's key.
This code can handle simple queries to hashes, complex value mappings
and repeated mappings for the same key.
Example:
my $sql = 'select key, one, two from table where something =?';
my $mapper = sub {
# Argument one is a row from the result, argument two is any previously seen value for the first column of the row
my ( $row, $value ) = @_;
#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 the default simplistic invocation
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";
# More complicated mapping, result hash will be keyed on "meta_key"
# Hash will contain lists of values linked to their meta_key
my %args = ( -SQL => 'select meta_key, meta_value from meta '
. 'where meta_key =? order by meta_id',
-PARAMS => ['species.classification'] );
my $hash = $helper->execute_into_hash(
%args,
-CALLBACK => sub {
my ( $row, $value ) = @_;
$value = [] if !defined $value;
push( @{$value}, $row->[1] );
return $value;
} );
# Add new values to an already seen existing meta_key
$hash = $helper->execute_into_hash(
%args,
-CALLBACK => sub {
my ( $row, $value ) = @_;
if ( defined $value ) {
# Calling code is dealing with $row->[0], so we only have to handle the remaining columns
push( @{$value}, $row->[1] );
return;
}
my $new_value = [ $row->[1] ];
return $new_value;
} );execute_single_result()
Arg [SQL] : string $sql
Arg [CALLBACK] : CodeRef The callback to use for mapping a row to a data point;
leave blank for a default scalar mapping
Arg [USE_HASHREFS] : boolean If set to true will cause HashRefs to be returned
to the callback & not ArrayRefs
Arg [PARAMS] : ArrayRef The binding parameters to the SQL statement
Arg [NO_ERROR] : Boolean Flag to indicate that the code should not
throw an error when row counts are not equal to 1
Returntype : Scalar
Exceptions : If errors occur in the execution of the SQL, if the query
returned more than 1 row and if we found no rows.
Status : Stable
Example :
my $meta_count =
$helper->execute_single_result(
-SQL => 'select count(*) from meta where species_id =?',
-PARAMS => [1] );
Description : Very similar to execute() except it will raise an exception if we have more
or less than one row returnedtransaction()
Arg [CALLBACK] : CodeRef The callback used for transaction isolation; once
the subroutine exists the code will decide on rollback
or commit. Required
Arg [RETRY] : integer the number of retries to attempt with this
transactional block. Defaults to 0.
Arg [PAUSE] : integer the time in seconds to pause in-between retries.
Defaults to 1. Fractions are allowed as use delegate to
Time::HiRes' sleep function
Arg [CONDITION] : CodeRef allows you to inspect the exception raised
and should your callback return true then the
retry will be attempted. If not given then all
exceptions mean attempt a retry (if specified)
Returntype : Return of the callback
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Example :
my $val = $helper->transaction(
-CALLBACK => sub {
my ($dbc) = @_;
#Do something
return 1;
} );
# Or without named arguments
my $val = $helper->transaction(
sub {
my ($dbc) = @_;
#Do something
return 1;
} );
# To allow retries (use sparingly, and see description)
my $val = $helper->transaction(
-RETRY => 3,
-PAUSE => 2,
-CALLBACK => sub {
my ($dbc) = @_;
#Do something
return 1;
} );
# Only retry when we have an error containing the phrase "deadlock"
my $val = $helper->transaction(
-RETRY => 3, -PAUSE => 2,
-CALLBACK => sub {
my ($dbc) = @_;
#Do something
return 1;
},
-CONDITION => sub {
my ($error) = @_;
return ( $error =~ /deadlock/ ) ? 1 : 0;
}
);
Description :
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 disconnect_when_idle() method to allow transactions to work as expected.
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 ensure the freshest data use SELECT ... FROM ... LOCK IN SHARE MODE> or SELECT ... FROM ... LOCK FOR UPDATE if you are going to issue updates.
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 and used).
You can also request the -RETRY of a transactional block of code which is causing problems. This can indicate your programming model is broken, so use with care.
The -RETRY argument indicates the number of times we attempt the transaction and -PAUSE indicates the time in-between attempts. These retries will only occur in the root transaction block i.e. you cannot influence the retry system in a sub-transaction. You can influence if the retry is done with the -CONDITION argument which accepts a Code reference (same as the -CALLBACK parameter). This allows you to inspect the error thrown to retry only in some situations e.g.
execute_update()
Arg [SQL] : string $sql
Arg [CALLBACK] : CodeRef The callback to use for calling methods on the
DBI statement handle or DBConnection object after an
update command
Arg [PARAMS] : ArrayRef The binding parameters to the SQL statement
Arg [PREPARE_PARAMS] : ArrayRef Parameters to bind to the prepare() StatementHandle call
Returntype : integer - Number of rows affected
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Description: Used for performing updates but conforms to the normal execute statement subroutines.
Example :
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 database 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 {
my ( $sth, $dbh, $rv ) = @_;
$obj->{id} = $dbh->{mysql_insertid};
},
-PARAMS => [ $obj->name() ] );#This lets us access the statement handle, the database handle and #the return value from $sth->execute, to access other properties such as #the last identifier inserted.
execute_with_sth()
Arg [SQL] : string $sql
Arg [CALLBACK] : CodeRef The callback to use for working with the statement
handle once returned. This is not a mapper.
Arg [PARAMS] : ArrayRef The binding parameters to the SQL statement
Arg [PREPARE_PARAMS] : ArrayRef Used to pass parameters to the statement handle
prepare method
Description : Run query without worrying statement handles and such.
Very similar to execute() except this gives you full control over the
lifecycle of the statement handle and 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 slowness only becomes apparent when working with very large numbers of rows. Returntype : Anything you wish to return from the callback Exceptions : If errors occur in the execution of the SQL Status : Stable Example : 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; } );
batch()
Arg [SQL] : string $sql
Arg [CALLBACK] : CodeRef The callback to use for working with the statement
handle once returned; specify this or -DATA
Arg [DATA] : ArrayRef 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.
Arg [PREPARE_PARAMS] : ArrayRef Used to pass parameters to the statement handle
prepare method
Returntype : integer rows updated
Exceptions : If errors occur in the execution of the SQL
Status : Stable
Example :
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 );
Description: Takes in a sql statement and a code reference. Your SQL is converted into a
prepared statement and 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 and 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 the same as 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 number affected rows by the query.