Spring Data JPA + JpaSpecificationExecutor + Entit

Question

(Using Spring Data JPA) I have two entities Parent& Child with a OneToMany/ManyToOne bi-directional relationship between them. I add a @NamedEntityGraph to the parent entity like so:

@Entity
@NamedEntityGraph(name = "Parent.Offspring", attributeNodes = @NamedAttributeNodes("children"))
public class Parent{
//blah blah blah

@OneToMany(mappedBy = "parent", fetch = FetchType.LAZY)
Set<Child> children;

//blah blah blah
}

Notice that the fetch type for the Parent's children is LAZY. This is on purpose. I don't always want to eager load the children when I'm querying an individual parent. Normally I could use my named entity graph to eager load the children on-demand, so to speak. But.....

There is a specific situation where I'd like to query for one or more parents AND eager load their children. In addition to this I need to be able to build this query programmatically. Spring Data provides the JpaSpecificationExecutor which allows one to build dynamic queries, but I can't figure out how to use it in conjunction with entity graphs for eager loading children in this specific case. Is this even possible? Is there some other way to eager load 'toMany entities using specifications?

Answer 1

Project Spring Data JPA EntityGraph implements some of the approaches mentioned in the other answers.

It has for example these additional repository interfaces:

• EntityGraphJpaRepository which is equivalent to standard JpaRepository
• EntityGraphJpaSpecificationExecutor which is equivalent to standard JpaSpecificationExecutor

Check the reference documentation for some examples.

Answer 2

The solution is to create a custom repository interface that implements these features:

@NoRepositoryBean
public interface CustomRepository<T, ID extends Serializable> extends JpaRepository<T, ID>, JpaSpecificationExecutor<T> {

    List<T> findAll(Specification<T> spec, EntityGraphType entityGraphType, String entityGraphName);
    Page<T> findAll(Specification<T> spec, Pageable pageable, EntityGraphType entityGraphType, String entityGraphName);
    List<T> findAll(Specification<T> spec, Sort sort, EntityGraphType entityGraphType, String entityGraphName);
    T findOne(Specification<T> spec, EntityGraphType entityGraphType, String entityGraphName);

}

Also create an implementation:

@NoRepositoryBean
public class CustomRepositoryImpl<T, ID extends Serializable> extends SimpleJpaRepository<T, ID> implements CustomRepository<T, ID> {

    private EntityManager em;

    public CustomRepositoryImpl(Class<T> domainClass, EntityManager em) {
        super(domainClass, em);
        this.em = em;
    }

    @Override
    public List<T> findAll(Specification<T> spec, EntityGraph.EntityGraphType entityGraphType, String entityGraphName) {
        TypedQuery<T> query = getQuery(spec, (Sort) null);
        query.setHint(entityGraphType.getKey(), em.getEntityGraph(entityGraphName));
        return query.getResultList();
    }

    @Override
    public Page<T> findAll(Specification<T> spec, Pageable pageable, EntityGraph.EntityGraphType entityGraphType, String entityGraphName) {
        TypedQuery<T> query = getQuery(spec, pageable.getSort());
        query.setHint(entityGraphType.getKey(), em.getEntityGraph(entityGraphName));
        return readPage(query, pageable, spec);
    }

    @Override
    public List<T> findAll(Specification<T> spec, Sort sort, EntityGraph.EntityGraphType entityGraphType, String entityGraphName) {
        TypedQuery<T> query = getQuery(spec, sort);
        query.setHint(entityGraphType.getKey(), em.getEntityGraph(entityGraphName));
        return query.getResultList();
    }

    @Override
    public T findOne(Specification<T> spec, EntityGraph.EntityGraphType entityGraphType, String entityGraphName) {
        TypedQuery<T> query = getQuery(spec, (Sort) null);
        query.setHint(entityGraphType.getKey(), em.getEntityGraph(entityGraphName));
        return query.getSingleResult();
    }
}

And create a factory:

public class CustomRepositoryFactoryBean<R extends JpaRepository<T, I>, T, I extends Serializable> extends JpaRepositoryFactoryBean<R, T, I> {

    protected RepositoryFactorySupport createRepositoryFactory(EntityManager entityManager) {
        return new CustomRepositoryFactory(entityManager);
    }

    private static class CustomRepositoryFactory<T, I extends Serializable> extends JpaRepositoryFactory {

        private EntityManager entityManager;

        public CustomRepositoryFactory(EntityManager entityManager) {
            super(entityManager);
            this.entityManager = entityManager;
        }

        protected Object getTargetRepository(RepositoryMetadata metadata) {
            return new CustomRepositoryImpl<T, I>((Class<T>) metadata.getDomainType(), entityManager);
        }

        protected Class<?> getRepositoryBaseClass(RepositoryMetadata metadata) {
            // The RepositoryMetadata can be safely ignored, it is used by the JpaRepositoryFactory
            //to check for QueryDslJpaRepository's which is out of scope.
            return CustomRepository.class;
        }
    }

}

And change the default repository factory bean to the new bean, e.g. in spring boot add this to the configuration:

@EnableJpaRepositories(
    basePackages = {"your.package"},
    repositoryFactoryBeanClass = CustomRepositoryFactoryBean.class
)

For more info about custom repositories: http://docs.spring.io/spring-data/jpa/docs/current/reference/html/#repositories.custom-behaviour-for-all-repositories

Answer 3

To complement the answers of Joep and pbo, I have to say that with new versions of Spring Data JPA, you will have to modify the constructor of CustomRepositoryImpl. Now the documentation says:

The class needs to have a constructor of the super class which the store-specific repository factory implementation is using. In case the repository base class has multiple constructors, override the one taking an EntityInformation plus a store specific infrastructure object (e.g. an EntityManager or a template class).

I use the following constructor:

public CustomRepositoryImpl(JpaEntityInformation<T,?> entityInformation, EntityManager em) {
    super(entityInformation, em);
    this.domainClass = entityInformation.getJavaType();
    this.em = em;
}

I've also added a private field to store the domain class:

private final Class<T> domainClass;

This allow me to get rid of the deprecated method readPage(javax.persistence.TypedQuery<T> query, Pageable pageable, @Nullable Specification<T> spec) and use instead:

@Override
public Page<T> findAll(Specification<T> spec, Pageable pageable, EntityGraph.EntityGraphType entityGraphType, String entityGraphName) {
    TypedQuery<T> query = getQuery(spec, pageable.getSort());
    query.setHint(entityGraphType.getKey(), em.getEntityGraph(entityGraphName));
    return readPage(query, domainClass, pageable, spec);
 }

Answer 4

The Joepie response is O.K.

but you don't need to create repositoryFactoryBeanClass, set up repositoryBaseClass

@EnableJpaRepositories(
    basePackages = {"your.package"},
    repositoryBaseClass = CustomRepositoryImpl.class)