问题
Lets consider these entities
@Entity
public class Room{
@Id
private Integer id;
private String number;
private String floor;
@ManyToOne
private RoomType roomType;
// Setters & Getters
}
@Entity
public class RoomType{
@Id
private Integer id;
private String name;
private String description;
private Boolean enabled;
// Setters & Getters
}
And also this interface for projection alongside repository class
public interface RoomList{
public Number getId();
public String getNumber();
public RoomType getRoomType();
interface RoomType {
String getName();
}
}
@Repository
public interface RoomRepository extends JpaRepository<Room,Integer>{
public Collection<RoomList> findAllProjectedBy();
}
Now if I look at generated SQL
select
room0_.id as col_0_0_,
room0_.number as col_1_0_,
roomtype1_.id as id1_3_,
roomtype1_.description as descript2_3_,
roomtype1_.enabled as isActive3_3_,
roomtype1_.name as name5_3_
from
Room room0_
inner join
roomType roomtype1_
on room0_.roomType_id=roomtype1_.id
The generated query should be something like this
select
room0_.id as col_0_0_,
room0_.number as col_1_0_,
roomtype1_.name as name5_3_
from
Room room0_
inner join
roomType roomtype1_
on room0_.roomType_id=roomtype1_.id
Can someone explain this behaviour or either this is a bug ? also what other options do we have achieve this kind of result. I already tried JPA entitygraph but graph type fetch is not yet fully supported in hibernate, i don't want to use constructor jpql query either. Thanks !
回答1:
If understand correctly the concern is that more attributes/columns get selected than necessary to fill the projections.
As I just described in issue DATAJPA-1218 this is just how projections work in Spring Data currently. The attributes of referenced entities do not get limited to those used in the projection.
来源:https://stackoverflow.com/questions/47159647/spring-jpa-nested-projection-generating-incorrect-query