DQL stands for **D**octrine **Q**uery **L**anguage and is an Object Query Language derivate that is very similar to the **H**ibernate **Q**uery **L**anguage (HQL) or the **J**ava **P**ersistence **Q**uery **L**anguage (JPQL).
In essence, DQL provides powerful querying capabilities over your object model. Imagine all your objects lying around in some storage (like an object database). When writing DQL queries, think about querying that storage to pick a certain subset of your objects.
> **CAUTION**
> A common mistake for beginners is to mistake DQL for being just some form of SQL
> and therefore trying to use table names and column names or join arbitrary tables
> together in a query. You need to think about DQL as a query language for your object
> model, not for your relational schema.
DQL is case in-sensitive, except for namespace, class and field names, which are case sensitive.
++ Types of DQL queries
DQL as a query language has SELECT, UPDATE and DELETE constructs that map to their corresponding
SQL statement types. INSERT statements are not allowed in DQL, because entities and their relations
have to be introduced into the persistence context through `EntityManager#persist()` to ensure
consistency of your object model.
DQL SELECT statements are a very powerful way of retrieving parts of your domain model that are
not accessible via associations. Additionally they allow to retrieve entities and their associations
in one single sql select statement which can make a huge difference in performance in contrast
to using several queries.
DQL UPDATE and DELETE statements offer a way to execute bulk changes on the entities of your
domain model. This is often necessary when you cannot load all the affected entities of a bulk
update into memory.
++ SELECT queries
+++ DQL SELECT clause
The select clause of a DQL query specifies what appears in the query result. The composition of all the expressions in the select clause also influences the nature of the query result.
Here is an example that selects all users with an age > 20:
[php]
$query = $em->createQuery('SELECT u FROM MyProject\Model\User u WHERE u.age > 20');
$users = $query->getResult();
Lets examine the query:
* `u` is a so called identification variable or alias that refers to the `MyProject\Model\User` class. By placing this alias in the SELECT clause we specify that we want all instances of the User class that are matched by this query appear in the query result.
* The FROM keyword is always followed by a fully-qualified class name which in turn is followed by an identification variable or alias for that class name. This class designates a root of our query from which we can navigate further via joins (explained later) and path expressions.
* The expression `u.age` in the WHERE clause is a path expression. Path expressions in DQL are easily identified by the use of the '.' operator that is used for constructing paths. The path expression `u.age` refers to the `age` field on the User class.
The result of this query would be a list of User objects where all users are older than 20.
The SELECT clause allows to specify both class identification variables that signal the hydration
of a complete entity class or just fields of the entity using the syntax `u.name`.
Combinations of both are also allowed and it is possible to wrap both fields and
identification values into aggregation and DQL functions. Numerical fields can
be part of computations using mathematical operations. See the sub-section on [DQL Functions, Aggregates and Operations](#dqlfn)
on more information.
+++ Joins
A SELECT query can contain joins. There are 2 types of JOINs: "Regular" Joins and "Fetch" Joins.
**Regular Joins**: Used to limit the results and/or compute aggregate values.
**Fetch Joins**: In addition to the uses of regular joins: Used to fetch related entities and include them in the hydrated result of a query.
There is no special DQL keyword that distinguishes a regular join from a fetch join. A join (be it an inner or outer join) becomes a "fetch join" as soon as fields of the joined entity appear in the SELECT part of the DQL query outside of an aggregate function. Otherwise its a "regular join".
Example:
Regular join of the address:
[php]
$query = $em->createQuery("SELECT u FROM User u JOIN u.address a WHERE a.city = 'Berlin'");
$users = $query->getResult();
Fetch join of the address:
[php]
$query = $em->createQuery("SELECT u, a FROM User u JOIN u.address a WHERE a.city = 'Berlin'");
$users = $query->getResult();
When Doctrine hydrates a query with fetch-join it returns the class in the FROM clause on the
root level of the result array. In the previous example an array of User instances is returned
and the address of each user is fetched and hydrated into the `User#address` variable. If you access
the address Doctrine does not need to lazy load the association with another query.
> **NOTE**
> Doctrine allows you to walk all the associations between all the objects in your domain model.
> Objects that were not already loaded from the database are replaced with lazy load proxy instances.
> Non-loaded Collections are also replaced by lazy-load instances that fetch all the contained objects upon
> first access. However relying on the lazy-load mechanism leads to many small queries executed
> against the database, which can significantly affect the performance of your application.
> **Fetch Joins** are the solution to hydrate most or all of the entities that you
> need in a single SELECT query.
+++ Named and Positional Parameters
DQL supports both named and positional parameters, however in contrast to many SQL dialects
positional parameters are specified with numbers, for example "?1", "?2" and so on.
Named parameters are specified with ":name1", ":name2" and so on.
+++ DQL SELECT Examples
This section contains a large set of DQL queries and some explanations of what is happening.
The actual result also depends on the hydration mode.
Hydrate all User entities:
[php]
$query = $em->createQuery('SELECT u FROM MyProject\Model\User u');
$users = $query->getResult(); // array of User objects
Retrieve the IDs of all CmsUsers:
[php]
$query = $em->createQuery('SELECT u.id FROM CmsUser u');
$ids = $query->getResult(); // array of CmsUser ids
Retrieve the IDs of all users that have written an article:
This section demonstrates how you can query inherited classes and what type of results
to expect.
+++ Single Table
[Single Table Inheritance](http://martinfowler.com/eaaCatalog/singleTableInheritance.html) is an inheritance mapping strategy where all classes of a hierarchy are mapped to a single database table. In order to distinguish which row represents which type in the hierarchy a so-called discriminator column is used.
First we need to setup an example set of entities to use. In this scenario it is a generic
First notice that the generated SQL to create the tables for these entities looks like
the following:
[sql]
CREATE TABLE Person (id INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL, name VARCHAR(50) NOT NULL, discr VARCHAR(255) NOT NULL, department VARCHAR(50) NOT NULL)
Now when persist a new `Employee` instance it will set the discriminator value for us
automatically:
[php]
$employee = new \Entities\Employee();
$employee->setName('test');
$employee->setDepartment('testing');
$em->persist($employee);
$em->flush();
Now lets run a simple query to retrieve the `Employee` we just created:
SELECT p0_.id AS id0, p0_.name AS name1, p0_.department AS department2, p0_.discr AS discr3 FROM Person p0_ WHERE (p0_.name = ?) AND p0_.discr IN ('employee')
[Class Table Inheritance](http://martinfowler.com/eaaCatalog/classTableInheritance.html) is an inheritance mapping strategy where each class in a hierarchy is mapped to several tables: its own table and the tables of all parent classes. The table of a child class is linked to the table of a parent class through a foreign key constraint.
Doctrine 2 implements this strategy through the use of a discriminator column in the topmost table of the hierarchy because this is the easiest way to achieve polymorphic queries with Class Table Inheritance.
The example for class table inheritance is the same as single table, you just need to
change the inheritance type from `SINGLE_TABLE` to `JOINED`:
Now take a look at the SQL which is generated to create the table, you'll notice some
differences:
[sql]
CREATE TABLE Person (id INT AUTO_INCREMENT NOT NULL, name VARCHAR(50) NOT NULL, discr VARCHAR(255) NOT NULL, PRIMARY KEY(id)) ENGINE = InnoDB;
CREATE TABLE Employee (id INT NOT NULL, department VARCHAR(50) NOT NULL, PRIMARY KEY(id)) ENGINE = InnoDB;
ALTER TABLE Employee ADD FOREIGN KEY (id) REFERENCES Person(id) ON DELETE CASCADE
* The data is split between two tables
* A foreign key exists between the two tables
Now if were to insert the same `Employee` as we did in the `SINGLE_TABLE` example and run
the same example query it will generate different SQL joining the `Person` information
automatically for you:
[sql]
SELECT p0_.id AS id0, p0_.name AS name1, e1_.department AS department2, p0_.discr AS discr3 FROM Employee e1_ INNER JOIN Person p0_ ON e1_.id = p0_.id WHERE p0_.name = ?
An instance of the `Doctrine\ORM\Query` class represents a DQL query. You create a Query instance be calling `EntityManager#createQuery($dql)`, passing the DQL query string. Alternatively you can create an empty `Query` instance and invoke `Query#setDql($dql)` afterwards. Here are some examples:
[php]
// $em instanceof EntityManager
// example1: passing a DQL string
$q = $em->createQuery('select u from MyProject\Model\User u');
// example2: usin setDql
$q = $em->createQuery();
$q->setDql('select u from MyProject\Model\User u');
+++ Query Result Formats
The format in which the result of a DQL SELECT query is returned can be influenced by a so-called `hydration mode`. A hydration mode specifies a particular way in which an SQL result set is transformed. Each hydration mode has its own dedicated method on the Query class. Here they are:
* `Query#getResult()`: Retrieves a collection of objects. The result is either a plain collection of objects (pure) or an array where the objects are nested in the result rows (mixed).
* `Query#getSingleResult()`: Retrieves a single object. If the result contains more than one object, an exception is thrown. The pure/mixed distinction does not apply.
* `Query#getArrayResult()`: Retrieves an array graph (a nested array) that is largely interchangeable with the object graph generated by `Query#getResultList()` for read-only purposes.
> **NOTE**
> An array graph can differ from the corresponding object graph in
> certain scenarios due to the difference of the identity semantics between arrays and
> objects.
* `Query#getScalarResult()`: Retrieves a flat/rectangular result set of scalar values that can contain duplicate data. The pure/mixed distinction does not apply.
* `Query#getSingleScalarResult()`: Retrieves a single scalar value from the result returned by the dbms. If the result contains more than a single scalar value, an exception is thrown. The pure/mixed distinction does not apply.
Instead of using these methods, you can alternatively use the general-purpose method `Query#execute(array $params = array(), $hydrationMode = Query::HYDRATE_OBJECT)`. Using this method you can directly supply the hydration mode as the second parameter via one of the Query constants. In fact, the methods mentioned earlier are just convenient shortcuts for the execute method. For example, the method `Query#getResultList()` internally invokes execute, passing in `Query::HYDRATE_OBJECT` as the hydration mode.
The use of the methods mentioned earlier is generally preferred as it leads to more concise code.
+++ Pure and Mixed Results
The nature of a result returned by a DQL SELECT query retrieved through `Query#getResult()` or `Query#getArrayResult()` can be of 2 forms: **pure** and **mixed**. In the previous simple examples, you already saw a "pure" query result, with only objects. By default, the result type is **pure** but **as soon as scalar values, such as aggregate values or other scalar values that do not belong to an entity, appear in the SELECT part of the DQL query, the result becomes mixed**. A mixed result has a different structure than a pure result in order to accomodate for the scalar values.
A pure result usually looks like this:
array
[0] => Object
[1] => Object
[2] => Object
...
A mixed result on the other hand has the following general structure:
array
array
[0] => Object
[1] => "some scalar string"
['count'] => 42
// ... more scalar values, either indexed numerically or with a name
array
[0] => Object
[1] => "some scalar string"
['count'] => 42
// ... more scalar values, either indexed numerically or with a name
To better understand mixed results, consider the following DQL query:
[sql]
SELECT u, UPPER(u.name) nameUpper FROM MyProject\Model\User u
This query makes use of the `UPPER` DQL function that returns a scalar value and because there is now a scalar value in the SELECT clause, we get a mixed result.
Here is how the result could look like:
array
array
[0] => User (Object)
['nameUpper'] => "Roman"
array
[0] => User (Object)
['nameUpper'] => "Jonathan"
...
And here is how you would access it in PHP code:
[php]
foreach ($results as $row) {
echo "Name: " . $row[0]->getName();
echo "Name UPPER: " . $row['nameUpper'];
}
You may have observed that in a mixed result, the object always ends up on index 0 of a result row.
+++ Hydration Modes
Each of the Hydration Modes makes assumptions about how the result is returned to user land. You should
know about all the details to make best use of the different result formats:
The constants for the different hydration modes are:
* Query::HYDRATE_OBJECT
* Query::HYDRATE_ARRAY
* Query::HYDRATE_SCALAR
* Query::HYDRATE_SINGLE_SCALAR
++++ Object Hydration
Object hydration hydrates the result set into the object graph:
[php]
$query = $em->createQuery('SELECT u FROM CmsUser u');
> You can set the Result Cache Driver globally on the `Doctrine\ORM\Configuration` instance
> so that it is passed to every `Query` and `NativeQuery` instance.
++++ Query Hints
You can pass hints to the query parser and hydrators by using the `AbstractQuery::setHint($name, $value)` method.
Currently there exist mostly internal query hints that are not be consumed in userland. However the following few hints
are to be used in userland:
* Query::HINT_FORCE_PARTIAL_LOAD - Allows to hydrate objects although not all their columns are fetched. This query
hint can be used to handle memory consumption problems with large result-sets that contain char or binary data.
Doctrine has no way of implicitly reloading this data. Partially loaded objects have to be passed to
`EntityManager::refresh()` if they are to be reloaded fully from the database.
* Query::HINT_REFRESH - This query is used internally by `EntityManager::refresh()` and can be used in userland aswell.
If you specify this hint and a query returns the data for an entity that is already managed by the UnitOfWork, the
fields of the existing entity will be refreshed. In normal operation a result-set that loads data of an already existing
entity is discarded in favour of the already existing entity.
* Query::HINT_CUSTOM_TREE_WALKERS - An array of additional `Doctrine\ORM\Query\TreeWalker` instances that are attached
to the DQL query parsing process.
++++ Query Cache (DQL Query Only)
Parsing a DQL query and converting it into an SQL query against the underlying database platform obviously has some overhead
in contrast to directly executing Native SQL queries. That is why there is a dedicated Query Cache for caching the
DQL parser results. In combination with the use of wildcards you can reduce the number of parsed queries in production
to zero.
The Query Cache Driver is passed from the `Doctrine\ORM\Configuration` instance to each `Doctrine\ORM\Query` instance
by default and is also enabled by default. This also means you don't regularly need to fiddle with the parameters
of the Query Cache, however if you do there are several methods to interact with it:
* `Query::setQueryCacheDriver($driver)` - Allows to set a Cache instance
* `Query::setQueryCacheLifeTime($seconds = 3600)` - Set lifetime of the query caching.
* `Query::expireQueryCache($bool)` - Enforce the expiring of the query cache if set to true.
* `Query::getExpireQueryCache()`
* `Query::getQueryCacheDriver()`
* `Query::getQueryCacheLifeTime()`
++++ First and Max Result Items (DQL Query Only)
You can limit the number of results returned from a DQL query aswell as specify the starting offset, Doctrine
then uses a strategy of manipulating the select query to return only the requested number of results:
* `Query::setMaxResults($maxResults)`
* `Query::setFirstResult($offset)`
> **NOTE**
> If your query contains a fetch-joined collection specifying the result limit methods are not working
> as you would expect. Set Max Results restricts the number of database result rows, however in the
> case of fetch-joined collections one root entity might appear in many rows, effectivly hydrating
> less than the specified number of results.
++ EBNF
The following context-free grammar, written in an EBNF variant, describes the Doctrine Query Language. You can consult this grammar whenever you are unsure about what is possible with DQL or what the correct syntax for a particular query should be.
+++ Document syntax:
* non-terminals begin with an upper case character
* terminals begin with a lower case character
* parentheses (...) are used for grouping
* square brackets [...] are used for defining an optional part, eg. zero or one time
* curly brackets {...} are used for repetion, eg. zero or more times
* double quotation marks "..." define a terminal string a vertical bar | represents an alternative
