NAME

Mojo::SQLite::Database - Database

SYNOPSIS

use Mojo::SQLite::Database;

my $db = Mojo::SQLite::Database->new(sqlite => $sql, dbh => $dbh);
$db->query('select * from foo')
  ->hashes->map(sub { $_->{bar} })->join("\n")->say;

DESCRIPTION

Mojo::SQLite::Database is a container for DBD::SQLite database handles used by Mojo::SQLite.

ATTRIBUTES

Mojo::SQLite::Database implements the following attributes.

dbh

my $dbh = $db->dbh;
$db     = $db->dbh($dbh);

DBD::SQLite database handle used for all queries.

# Use DBI utility methods
my $quoted = $db->dbh->quote_identifier('foo.bar');

results_class

my $class = $db->results_class;
$db       = $db->results_class('MyApp::Results');

Class to be used by "query", defaults to Mojo::SQLite::Results. Note that this class needs to have already been loaded before "query" is called.

sqlite

my $sql = $db->sqlite;
$db     = $db->sqlite(Mojo::SQLite->new);

Mojo::SQLite object this database belongs to.

METHODS

Mojo::SQLite::Database inherits all methods from Mojo::Base and implements the following new ones.

begin

my $tx = $db->begin;
my $tx = $db->begin('exclusive');

Begin transaction and return Mojo::SQLite::Transaction object, which will automatically roll back the transaction unless "commit" in Mojo::SQLite::Transaction has been called before it is destroyed.

# Insert rows in a transaction
eval {
  my $tx = $db->begin;
  $db->insert('frameworks', {name => 'Catalyst'});
  $db->insert('frameworks', {name => 'Mojolicious'});
  $tx->commit;
};
say $@ if $@;

A transaction locking behavior of deferred, immediate, or exclusive may optionally be passed; the default in DBD::SQLite is currently immediate. See "Transaction and Database Locking" in DBD::SQLite and https://sqlite.org/lang_transaction.html for more details.

delete

my $results = $db->delete($table, \%where);

Generate a DELETE statement with "abstract" in Mojo::SQLite (usually an SQL::Abstract object) and execute it with "query". You can also append a callback for API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->delete(some_table => sub ($db, $err, $results) {
  ...
});
Mojo::IOLoop->start unless Mojo::IOLoop->is_running;

Use all the same argument variations you would pass to the delete method of SQL::Abstract.

# "delete from some_table"
$db->delete('some_table');

# "delete from some_table where foo = 'bar'"
$db->delete('some_table', {foo => 'bar'});

# "delete from some_table where foo like '%test%'"
$db->delete('some_table', {foo => {-like => '%test%'}});

delete_p

my $promise = $db->delete_p($table, \%where, \%options);

Same as "delete" but returns a Mojo::Promise object instead of accepting a callback. For API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->delete_p('some_table')->then(sub ($results) {
  ...
})->catch(sub ($err) {
  ...
})->wait;

disconnect

$db->disconnect;

Disconnect "dbh" and prevent it from getting reused.

insert

my $results = $db->insert($table, \@values || \%fieldvals, \%options);

Generate an INSERT statement with "abstract" in Mojo::SQLite (usually an SQL::Abstract object) and execute it with "query". You can also append a callback for API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->insert(some_table => {foo => 'bar'} => sub ($db, $err, $results) {
  ...
});
Mojo::IOLoop->start unless Mojo::IOLoop->is_running;

Use all the same argument variations you would pass to the insert method of SQL::Abstract.

# "insert into some_table (foo, baz) values ('bar', 'yada')"
$db->insert('some_table', {foo => 'bar', baz => 'yada'});

insert_p

my $promise = $db->insert_p($table, \@values || \%fieldvals, \%options);

Same as "insert" but returns a Mojo::Promise object instead of accepting a callback. For API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->insert_p(some_table => {foo => 'bar'})->then(sub ($results) {
  ...
})->catch(sub ($err) {
  ...
})->wait;

ping

my $bool = $db->ping;

Check database connection.

query

my $results = $db->query('select * from foo');
my $results = $db->query('insert into foo values (?, ?, ?)', @values);
my $results = $db->query('select ? as img', {type => SQL_BLOB, value => slurp 'img.jpg'});
my $results = $db->query('select ? as foo', {json => {bar => 'baz'}});

Execute a blocking SQL statement and return a results object based on "results_class" (which is usually Mojo::SQLite::Results) with the query results. The DBD::SQLite statement handle will be automatically reused when it is not active anymore, to increase the performance of future queries. You can also append a callback for API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->query('insert into foo values (?, ?, ?)' => @values => sub ($db, $err, $results) {
  ...
});
Mojo::IOLoop->start unless Mojo::IOLoop->is_running;

Hash reference arguments containing type and value elements will use the specified bind type for the parameter, using types from "DBI Constants" in DBI; see "Blobs" in DBD::SQLite and the subsequent section for more information.

Hash reference arguments containing a value named json or -json will be encoded to JSON text with "to_json" in Mojo::JSON. To accomplish the reverse, you can use the method "expand" in Mojo::SQLite::Results to decode JSON text fields to Perl values with "from_json" in Mojo::JSON.

# "I ♥ SQLite!"
$db->query('select ? as foo', {json => {bar => 'I ♥ SQLite!'}})
  ->expand(json => 'foo')->hash->{foo}{bar};

query_p

my $promise = $db->query_p('SELECT * FROM foo');

Same as "query" but returns a Mojo::Promise object instead of accepting a callback. For API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->query_p('INSERT INTO foo VALUES (?, ?, ?)' => @values)->then(sub ($results) {
  ...
})->catch(sub ($err) {
  ...
})->wait;

select

my $results = $db->select($source, $fields, $where, $order);

Generate a SELECT statement with "abstract" in Mojo::SQLite (usually an SQL::Abstract object) and execute it with "query". You can also append a callback for API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->select(some_table => ['foo'] => {bar => 'yada'} => sub ($db, $err, $results) {
  ...
});
Mojo::IOLoop->start unless Mojo::IOLoop->is_running;

Use all the same argument variations you would pass to the select method of SQL::Abstract.

# "select * from some_table"
$db->select('some_table');

# "select id, foo from some_table"
$db->select('some_table', ['id', 'foo']);

# "select * from some_table where foo = 'bar'"
$db->select('some_table', undef, {foo => 'bar'});

# "select * from some_table where foo = 'bar' order by id desc"
$db->select('some_table', undef, {foo => 'bar'}, {-desc => 'id'});

# "select * from some_table where foo like '%test%'"
$db->select('some_table', undef, {foo => {-like => '%test%'}});

select_p

my $promise = $db->select_p($source, $fields, $where, \%options);

Same as "select" but returns a Mojo::Promise object instead of accepting a callback. For API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->select_p(some_table => ['foo'] => {bar => 'yada'})->then(sub ($results) {
  ...
})->catch(sub ($err) {
  ...
})->wait;

tables

my $tables = $db->tables;

Return table and view names for this database, that are visible to the current user and not internal, as an array reference. Names will be quoted and prefixed by a schema name of "main" for standard tables, "temp" for temporary tables, and the appropriate schema name for attached databases.

# Names of all tables
say for @{$db->tables};

update

my $results = $db->update($table, \%fieldvals, \%where);

Generate an UPDATE statement with "abstract" in Mojo::SQLite (usually an SQL::Abstract object) and execute it with "query". You can also append a callback for API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->update(some_table => {foo => 'baz'} => {foo => 'bar'} => sub ($db, $err, $results) {
  ...
});
Mojo::IOLoop->start unless Mojo::IOLoop->is_running;

Use all the same argument variations you would pass to the update method of SQL::Abstract.

# "update some_table set foo = 'bar' where id = 23"
$db->update('some_table', {foo => 'bar'}, {id => 23});

# "update some_table set foo = 'bar' where foo like '%test%'"
$db->update('some_table', {foo => 'bar'}, {foo => {-like => '%test%'}});

update_p

my $promise = $db->update_p($table, \%fieldvals, \%where, \%options);

Same as "update" but returns a Mojo::Promise object instead of accepting a callback. For API compatibility with Mojo::Pg; the query is still executed in a blocking manner.

$db->update_p(some_table => {foo => 'baz'} => {foo => 'bar'})->then(sub ($results) {
  ...
})->catch(sub ($err) {
  ...
})->wait;

BUGS

Report any issues on the public bugtracker.

AUTHOR

Dan Book, dbook@cpan.org

COPYRIGHT AND LICENSE

Copyright 2015, Dan Book.

This library is free software; you may redistribute it and/or modify it under the terms of the Artistic License version 2.0.

SEE ALSO

Mojo::SQLite