Doctrine ORM HYDRATE_SCALAR vs HYDRATE_ARRAY - doctrine-orm

Question

I'm trying to get all values of a column from a table through Doctrine ORM using following query:
SELECT item.templateId FROM MyBundle:Item item WHERE item.id IN (:ids)
I know Doctrine has hydration mode Query::HYDRATE_SINGLE_SCALAR which returns single scalar value. When, however, having more than one record in result it throws NonUniqueResultException which is fine.
Naturally, I guessed that Query::HYDRATE_SCALAR would do the trick. However, it returns exactly the same result as when using Query::HYDRATE_ARRAY.
Code:
$ids = [1, 2, 3];
$query = $this->entityManager->createQuery('
SELECT item.templateId FROM MyBundle:Item item WHERE item.id IN (:ids)
')
->setParameter('ids', $ids)
;
// Regardles whether I use HYDRATE_SCALAR or HYDRATE_ARRAY
// $result is the same
$result = $query->getResult(Query::HYDRATE_SCALAR);
Here is how $result looks in both cases:
Array
(
[0] => Array
(
[templateId] => 52
)
[1] => Array
(
[templateId] => 90
)
[2] => Array
(
[templateId] => 89
)
)
How I wish it looked:
Array
(
[0] => 52
[1] => 90
[2] => 89
)
I know I can transform it into the form I want, but I wonder why HYDRATE_SCALAR isn't behaving as I expect it to.

Answer 1

I think it behaves like this because, as the documentation says :
[it] can contain duplicate data.
That's when array_column (php 5.5) comes into play.

Answer 2

Try this:
$result = $query->getResult(Query::HYDRATE_SCALAR);
return array_map('reset', $result);
There you go :)

Answer 3

10 years later, I'm adding another answer, that I will hope is better.
Once again, the documentation I mentioned 10 years ago is helpful:
Query#getArrayResult(): Retrieves an array graph (a nested array) that is largely interchangeable with the object graph generated by Query#getResult() 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.
ORM means Object-relational mapper: it maps an RDBMS with an object graph. One one hand you have tabular data, on the other hand you have objects that can be related to each other.
When using getArrayResult(), you will get an array graph, meaning something (deeply) nested, like a JSON document.
When using getScalarResult(), there will be little to no intervention besides converting column names to field names, and that's why it can contain duplicate data. So for instance, If you write a query that retrieves a user with 2 phonenumbers, the former will give you something like
[0 => [
'lastname' => 'Anderson',
'firstname' => 'Thomas',
'phoneNumbers' => [[
'alias' => 'home',
'number' => '0123456789',
], [
'alias' => 'work',
'number' => '987654321',
]],
]]
The latter will give you something flat and duplicated, like this:
[0 => [
'lastname' => 'Anderson',
'firstname' => 'Thomas',
'alias' => 'work',
'number' => '987654321',
],
],
1 => [
'lastname' => 'Anderson',
'firstname' => 'Thomas',
'alias' => 'work',
'number' => '987654321',
]]
Of course, if your query is simple enough, there is no difference.