Mapping 1-1 relationship in Hibernate ORM

Update notes:

30/08/2025: add use case for 1-1 mapping in RDBMS.

Context

Hibernate uses @OneToOne annotation to model 1-1 relationship in RDBMS. When we want to travel through entity relationship, we model these relationships as bidirectional ones. But with one-to-one relationship, hibernate will trigger additional query to fetch child entity if we map the relationship like others relationship.
In hibernate context, an object will be identified by id and entity type (aka class) [1].

Solution

When we model relationship as regular bidirectional one-to-one relationship like two entity below:

public class QuizEntity {

  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;

  @OneToOne(fetch = LAZY, mappedBy = "quiz", optional = false)
  private QuizResultEntity result;

  // other fields
}

public class QuizResultEntity {

  @Id
  @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;

  @OneToOne(fetch = LAZY, optional = false)
  @JoinColumn(name = "quiz_id")
  private QuizEntity quiz;

  // other fields
}

and fetch an entity of QuizEntity type, hibernate will trigger an additional query to check whether the associated QuizResultEntity object existing or not to provide appropriate strategy to initialize relationship. If QuizResultEntity object exists, hibernate will create a proxy, otherwise this field is set to null.

So if we add optional = false properties to @OneToOne annotation, will hibernate trigger additional query?

Unfortunately, based on [1], hibernate will still trigger additional query because the identifier of QuizResultEntity is unknown.

Query log:

select q1_0.id from quiz q1_0 where q1_0.id=?
select q1_0.id,q1_0.quiz_id from quiz_result q1_0 where q1_0.quiz_id=?

-> Use @OneToOne mapping with @MapIds at QuizResultEntity side can solve this problem. This will work because at QuizEntity side, hibernate knows the ID of associated QuizResult entity when it queries data.

public class QuizResultEntity {

  @Id
  // Note: we don't use auto generate strategy here.
  // @GeneratedValue(strategy = GenerationType.IDENTITY)
  private Long id;

  @OneToOne(fetch = LAZY)
  @MapsId
  @JoinColumn(name = "quiz_id")
  private QuizEntity quiz;

  // other fields
}

Query log:

select q1_0.id from quiz q1_0 where q1_0.id=?

Use case

With these complex mappings, why on earth would you want to use them? Of course, it does not exist for fun. There are many cases in which one-to-one mappings are useful.

Vertical partition: when an entity has many fields, we can split it into multiple tables based on query pattern to improve performance. For example, we have a User entity with many fields, we can split it into User and UserProfile entities. The UserProfile entity will have a 1-1 relationship with the User entity.
- The user_id will be the mapping key.
- These tables may reside in separate physical databases.
Avoid NULLs and subtype discrimination: when an entity has multiple subtypes, for example, we have a Product entity with subtypes Software_Product and Hardware_Product, we can use 1-1 mapping to store the common fields in the Product table and the subtype-specific fields in separate tables.

Appendix

Example source code: https://github.com/dntam00/learning-spring/tree/master/one-to-one

Tags: java hibernate