Magic-query is a PHP library that helps you work with complex SQL queries.
It comes with 3 great features:
Simply use the composer package:
{
"require": {
"mouf/magic-query": "^1.2"
},
"minimum-stability": "dev",
"prefer-stable": true
}
Just write the query with all possible parameters.
use Mouf\Database\MagicQuery;
$sql = "SELECT * FROM users WHERE name LIKE :name AND country LIKE :country";
// Get a MagicQuery object.
$magicQuery = new MagicQuery();
// Let's pass only the "name" parameter
$result = $magicQuery->build($sql, [ "name" => "%John%" ]);
// $result = SELECT * FROM users WHERE name LIKE '%John%'
// Did you notice how the bit about the country simply vanished?
// Let's pass no parameter at all!
$result2 = $magicQuery->build($sql, []);
// $result2 = SELECT * FROM users
// The whole WHERE condition disappeared because it is not needed anymore!
Curious to know how this work? Check out the parameters guide!
Fed up of writing joins in SQL? Let MagicQuery do the work for you!
Seriously? Yes! All you have to do is:
magicjoin(start_table)
Let's assume your database schema is:
Using MagicJoin, you can write this SQL query:
SELECT users.* FROM MAGICJOIN(users) WHERE groups.name = 'Admins' AND country.name='France';
and it will automatically be transformed into this:
SELECT users.* FROM users
LEFT JOIN users_groups ON users.user_id = users_groups.user_id
LEFT JOIN groups ON groups.group_id = users_groups.group_id
LEFT JOIN country ON country.country_id = users.country_id
WHERE groups.name = 'Admins' AND country.name='France';
And the code is so simple!
use Mouf\Database\MagicQuery;
$sql = "SELECT users.* FROM MAGICJOIN(users) WHERE groups.name = 'Admins' AND country.name='France'";
// Get a MagicQuery object.
// $conn is a Doctrine DBAL connection.
$magicQuery = new MagicQuery($conn);
$completeSql = $magicQuery->build($sql);
// $completeSql contains the complete SQL request, with all joins.
Want to know more? Check out the MagicJoin guide!
Discarding unused parameters and auto-joining keys is not enough? You have very specific needs? Say hello to Twig integration!
Using Twig integration, you can directly add Twig conditions right into your SQL.
use Mouf\Database\MagicQuery;
$sql = "SELECT users.* FROM users {% if isAdmin %} WHERE users.admin = 1 {% endif %}";
$magicQuery = new MagicQuery();
// By default, Twig integration is disabled. You need to enable it.
$magicQuery->setEnableTwig(true);
$completeSql = $magicQuery->build($sql, ['isAdmin' => true]);
// Parameters are passed to the Twig SQL query, and the SQL query is returned.
{{ id }}
, you should write :id
.Want to know more? Check out the MagicTwig guide!
No. By default, your SQL is parsed and then rewritten using the MySQL dialect, but you use any kind of dialect
known by Doctrine DBAL. Magic-query optionally uses Doctrine DBAL. You can pass a Connection
object
as the first parameter of the MagicQuery
constructor. Magic-query will then use the matching dialect.
For instance:
$config = new \Doctrine\DBAL\Configuration();
$connectionParams = array(
'url' => 'sqlite:///somedb.sqlite',
);
$conn = \Doctrine\DBAL\DriverManager::getConnection($connectionParams, $config);
$magicQuery = new \Mouf\Database\MagicQuery($conn);
Also, if you have no connection to your database configured but you want to generate SQL in some specific dialect, you can instead set the DBAL database platform used:
$magicQuery->setOutputDialect(new \Doctrine\DBAL\Platforms\PostgreSqlPlatform());
$magicQuery = new \Mouf\Database\MagicQuery();
MagicQuery does a lot to your query. It will parse it, render it internally as a tree of SQL nodes, etc... This processing is time consuming. So you should definitely consider using a cache system. MagicQuery is compatible with Doctrine Cache. You simply have to pass a Doctrine Cache instance has the second parameter of the constructor.
use Mouf\Database\MagicQuery;
use Doctrine\Common\Cache\ApcCache;
// $conn is a Doctrine connection
$magicQuery = new MagicQuery($conn, new ApcCache());
With MagicQuery, a lot happens to your SQL query. In particular, it is parsed using a modified version of the php-sql-parser library. If you face any issues with a complex query, it is likely there is a bug in the parser. Please open an issue on Github and we'll try to fix it.