Lightweight database abstraction in which rows are simple arrays. It has both a query builder for convenient fluent syntax and an interface for performing pure SQL queries.
$database = Database::create([
'driver' => 'mysql',
'dsn' => 'mysql:host=localhost;dbname=my_database',
'username' => 'root',
'password' => 'qwerty',
'prefix' => 'test_'
]);
$database->statement('
CREATE TABLE '.$database->addTablePrefix('users').' (
id INT(11) NOT NULL AUTO_INCREMENT,
email VARCHAR(50) NOT NULL,
account INT(11) NOT NULL DEFAULT 0
)
');
$database->table('users')->insert([
['name' => 'Jack', 'account' => 1200],
['name' => 'Bob', 'account' => 500],
['name' => 'Richard', 'account' => 800]
]);
$database->table('users')->where('account', '>', 600)->get(); // Jack and Richard
Key features:
- Light with a small number of light dependencies.
- Extensible. Examples will come soon.
- The query builder and the database connector can be used separately.
- Supports table prefixes.
- No static facades. Explicit delivery using dependency injection.
- Exceptions on errors.
Supported DBMSs:
- MySQL
- SQLite
- Maybe any other, didn't test it
If you need a new database system support please implement it there and there using pull requests.
You need Composer to use this library. Run in a console:
composer require finesse/mini-db
You need to make a Database
instance once:
use Finesse\MiniDB\Database;
$database = Database::create([
'driver' => 'mysql', // DBMS type: 'mysql', 'sqlite' or anything else for other (optional)
'dsn' => 'mysql:host=host;dbname=db', // PDO data source name (DSN)
'username' => 'root', // Database username (optional)
'password' => 'qwerty', // Database password (optional)
'options' => [], // PDO options (optional)
'prefix' => '' // Tables prefix (optional)
]);
See more about the PDO options at the PDO constructor reference.
Alternatively you can create all the dependencies manually:
use Finesse\MicroDB\Connection;
use Finesse\MiniDB\Database;
use Finesse\QueryScribe\Grammars\MySQLGrammar;
use Finesse\QueryScribe\PostProcessors\TablePrefixer;
$connection = Connection::create('mysql:host=host;dbname=db', 'username', 'password');
$grammar = new MySQLGrammar();
$tablePrefixer = new TablePrefixer('demo_');
$database = new Database($connection, $grammar, $tablePrefixer);
$database->insertGetId('INSERT INTO users (name, email) VALUES (?, ?), (?, ?)', ['Ann', 'ann@gmail.com', 'Bob', 'bob@rambler.com']); // 19 (the last inserted row id)
$database->select('SELECT * FROM users WHERE name = ? OR email = ?', ['Jack', 'jack@example.com']);
/*
[
['id' => 4, 'name' => 'Jack', 'email' => 'demon@mail.com', 'account' => 1230],
['id' => 17, 'name' => 'Bill', 'email' => 'jack@example.com', 'account' => -100]
]
*/
$database->import('path/to/file.sql');
The cell values are returned as they are returned by the underlying database connection. They are not casted automatically because casting can cause a data loss.
Table prefix is not applied in raw queries. Use $database->addTablePrefix()
to apply it.
$database->select('SELECT * FROM '.$database->addTablePrefix('users').' ORDER BY id');
Be careful, the statements
and the import
methods don't throw an exception if the second or a next statement of the
query has an error. This is a PDO bug.
You can find more information and examples of raw queries there.
Basic examples are presented here. You can find more cool examples there.
Values given to the query builder are treated safely to prevent SQL injections so you don't need to escape them.
Many rows:
$database
->table('users')
->where('status', 'active')
->orderBy('name')
->offset(40)
->limit(10)
->get();
/*
[
['id' => 17, 'name' => 'Bill', 'email' => 'jack@example.com', 'status' => 'active'],
['id' => 4, 'name' => 'Jack', 'email' => 'demon@mail.com', 'status' => 'active']
]
*/
One row:
$database
->table('users')
->where('status', 'active')
->orderBy('name')
->first();
/*
['id' => 17, 'name' => 'Bill', 'email' => 'jack@example.com', 'status' => 'active'] or null
*/
We suggest Pagerfanta to make a pagination easily.
First install Pagerfanta using composer by running in a console:
composer require pagerfanta/pagerfanta
Then make a query from which the rows should be taken:
$query = $database
->table('posts')
->where('category', 'archive')
->orderBy('date', 'desc');
// Don't call ->get() here
And use Pagerfanta:
use Finesse\MiniDB\ThirdParty\PagerfantaAdapter;
use Pagerfanta\Pagerfanta;
$paginator = new Pagerfanta(new PagerfantaAdapter($query));
$paginator->setMaxPerPage(10); // The number of rows on a page
$paginator->setCurrentPage(3); // The current page number
$currentPageRows = $paginator->getCurrentPageResults(); // The rows for the current page
$pagesCount = $paginator->getNbPages(); // Total pages count
$haveToPaginate = $paginator->haveToPaginate(); // Whether the number of results is higher than the max per page
You can find more reference and examples for Pagerfanta there.
If you need to process a large amount of rows you can use chunking. In this approach portions of rows are fetched from the database instead of fetching all the rows at once.
$database
->table('users')
->orderBy('id')
->chunk(100, function ($users) {
foreach ($users as $user) {
// Process a row here
}
});
$database
->table('products')
->where('price', '>', 1000)
->count(); // 31
Other aggregate methods: avg(column)
, sum(column)
, min(column)
and max(column)
.
Many rows:
$database->table('debts')->insert([
['name' => 'Sparrow', 'amount' => 13000, 'message' => 'Sneaky guy'],
['name' => 'Barbos', 'amount' => 4999, 'message' => null],
['name' => 'Pillower', 'message' => 'Call tomorrow']
]); // 3 (number of inserted rows)
The string array keys are the columns names.
One row:
$database->table('debts')->insertGetId([
'name' => 'Bigbigger',
'amount' => -3500,
'message' => 'I owe him'
]); // 4 (id of the inserted row)
From a select query:
$database->table('debts')->insertFromSelect(['name', 'amount', 'message'], function ($query) {
$query
->from('users')
->addSelect(['name', $query->raw('- account'), 'description'])
->where('status', 'debtor');
}); // 6 (number of inserted rows)
$database
->table('posts')
->where('date', '<', '2017-01-01')
->update([
'status' => 'obsolete',
'category' => null
]); // 5 (number of updated rows)
The array keys are the columns names.
$database
->table('messages')
->where('sender_id', 1456)
->orWhere('status', 'stink')
->delete(); // 5 (number of deleted rows)
Escape LIKE special wildcard characters:
$searchString = '%iamhacker%';
$query->where('name', 'like', $query->escapeLikeWildcards($searchString).'_'); // "name" LIKE \%iamhacker\%_
Wrap a table or a column name in quotes:
$query->whereRaw('MIN('.$query->quoteIdentifier('data"base').'.'.$query->quoteIdentifier('ta"ble').') > 10');
// or
$query->whereRaw('MIN('.$query->quoteCompositeIdentifier('data"base.ta"ble').') > 10'); // MIN("data""base"."ta""ble") > 10
The above methods are also available in a Database
object.
Make all the column names in the query have explicit table name or alias:
$database
->table('users', 'u')
->addSelect('name')
->where('status', 'verified')
->orWhere('u.type', 'admin')
->addTablesToColumnNames();
// SELECT "name" FROM "users" AS "u" WHERE "status" = ? OR "u"."type" = ?
The project follows the Semantic Versioning.
MIT. See the LICENSE file for details.