A query load and hydrate a main entity with it's joined entities, which are filtered.
Same query with a different filtering did not update the joined entities.
Some data:
Tiers:
| id | name |
| 1 | alpha |
| 2 | beta |
Container:
| id | tiers_id | category |
| 10 | 1 | A |
| 20 | 1 | A |
| 30 | 1 | B |
| 40 | 1 | B |
Execute 2 queries to get some tiers with theirs containers joined, category A first, then category B:
$dql = "select t, c
from Tiers t
join t.containers c
where t.id in (?1) and c.category = (?2)";
$result = $em->createQuery($dql)
->setParameter(1, array(1))
->setParameter(2, 'A')
->getResult();
$tiers = $result[0];
$containers = $tiers->getContainers(); // tiers 1 with containers 10 and 20, that's fine !
$result = $em->createQuery($dql)
->setParameter(1, array(1))
->setParameter(2, 'B')
->getResult();
$tiers = $result[0];
$containers = $tiers->getContainers(); // BAD HERE: still get containers 10 and 20, looking for containers 30 and 40.
After the 2nd query, tiers 1 preserve its containers loaded during first query. That's not what is expected.
So is there a way to get the containers 30 and 40 after 2nd query ?
Maybe a kind of "reset/detach" the containers of tiers entities after the first query ?
Or anything else...
Mutiple select in queries are used to hydrate tiers with required containers joined.
'getContainers' method gives the expected containers from each tiers.
And cost to BDD is only 1 SQL query, whatever the quantity of tiers searched.
I suppose tiers cannot be detach/reload because they are updated before, between and after the queries, it throw this kind of exception when flushing:
Uncaught Exception: Multiple non-persisted new entities were found through the given association graph
* A new entity was found through the relationship 'XXX' that was not configured to cascade persist operations for entity: XXX\Entity\Tiers@00000000257b87500000000018499b62.