Hibernate Spring数据-在@OneToOne上进行即时获取

我有两个实体：

@实体 @Tablename=某物 公共类SomethingEntity实现可序列化{ @身份证 @GeneratedValuestrategy=GenerationType.SEQUENCE，generator=SOMETHING_SEQ @SequenceGeneratorsequenceName=SOMETHING，allocationSize=1，name=SOMETHING @Columnname=SOMETHING\u ID，精度=18，比例=0，可空=false 私人长某物； @Columnname=SIMPLE_字段，长度=3，可空=true 私有字符串simpleField； @OneToOnecascade=CascadeType.ALL，mappedBy=somethingId 私人或其他实体； } @实体 @Tablename=OTHER 公共类OtherEntity实现可序列化{ @身份证 @奥内托内 @JoinColumnname=SOMETHING\u ID 私人某物实体某物ID； @Columnname=OTHER\u SIMPLE\u字段，精度=18，比例=2，可空=false 私人双基地； } 我希望Spring数据能够获取其他实体。我知道默认情况下它不起作用，所以我的存储库方法是：

@EntityGraph(attributePaths = {"other"})
List<SomethingEntity> findAll(Predicate predicate);

它不起作用，我猜是因为@OneToOne。我还尝试了连接获取，在我的示例中，simpleField谓词的简单选择是QueryDSL，没有任何效果

---

最令人惊讶的是，初始SQL实际上是正确的——它包含正确的连接并选择正确的字段。然后Spring数据按id加载每个@OneToOne关系，导致100个查询而不是1个查询。

我不理解您对@EntityGraph的使用，请尝试在实体中使用它，如下所示：

@NamedEntityGraphs({
    @NamedEntityGraph(name = "graph.SomethingEntity.OtherEntity", attributeNodes = @NamedAttributeNode("other"))
})
public class SomethingEntity implements Serializable {
[...]

EntityGraph graph = this.em.getEntityGraph("graph.SomethingEntity.OtherEntity");
Map hints = new HashMap();
hints.put("javax.persistence.fetchgraph", graph);//Important fetchgraph no other type
return entityManager.find(SomethingEntity.class,ID_OF_SOMETHING_ENTITY,  hints);

CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<SomethingEntity> cq = cb.createQuery(SomethingEntity.class);

//Need to define variable root to apply fetch
Root<SomethingEntity> rootSE = cq.from(SomethingEntity.class);

/* Use metamodel (in eclipse, property project / JPA / Canonical metamodel
 * and select src folder in source to allow autogenerated metamodels)
 */
rootSE.fetch(SomethingEntity_.other);

// Set your predicates
cq.where(cb.equals(rootSE.get(SomethingEntity_.somethingId),ID_OF_SOMETHING_ENTITY));

cq.select(rootSE);

entityManager.createQuery(cq).getSingleResult();

并在存储库方法中使用实体图，如下所示：

@NamedEntityGraphs({
    @NamedEntityGraph(name = "graph.SomethingEntity.OtherEntity", attributeNodes = @NamedAttributeNode("other"))
})
public class SomethingEntity implements Serializable {
[...]

EntityGraph graph = this.em.getEntityGraph("graph.SomethingEntity.OtherEntity");
Map hints = new HashMap();
hints.put("javax.persistence.fetchgraph", graph);//Important fetchgraph no other type
return entityManager.find(SomethingEntity.class,ID_OF_SOMETHING_ENTITY,  hints);

CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<SomethingEntity> cq = cb.createQuery(SomethingEntity.class);

//Need to define variable root to apply fetch
Root<SomethingEntity> rootSE = cq.from(SomethingEntity.class);

/* Use metamodel (in eclipse, property project / JPA / Canonical metamodel
 * and select src folder in source to allow autogenerated metamodels)
 */
rootSE.fetch(SomethingEntity_.other);

// Set your predicates
cq.where(cb.equals(rootSE.get(SomethingEntity_.somethingId),ID_OF_SOMETHING_ENTITY));

cq.select(rootSE);

entityManager.createQuery(cq).getSingleResult();

另一个选项是在yout repository方法中创建CriteriaQuery，如下所示：

@NamedEntityGraphs({
    @NamedEntityGraph(name = "graph.SomethingEntity.OtherEntity", attributeNodes = @NamedAttributeNode("other"))
})
public class SomethingEntity implements Serializable {
[...]

EntityGraph graph = this.em.getEntityGraph("graph.SomethingEntity.OtherEntity");
Map hints = new HashMap();
hints.put("javax.persistence.fetchgraph", graph);//Important fetchgraph no other type
return entityManager.find(SomethingEntity.class,ID_OF_SOMETHING_ENTITY,  hints);

CriteriaBuilder cb = entityManager.getCriteriaBuilder();
CriteriaQuery<SomethingEntity> cq = cb.createQuery(SomethingEntity.class);

//Need to define variable root to apply fetch
Root<SomethingEntity> rootSE = cq.from(SomethingEntity.class);

/* Use metamodel (in eclipse, property project / JPA / Canonical metamodel
 * and select src folder in source to allow autogenerated metamodels)
 */
rootSE.fetch(SomethingEntity_.other);

// Set your predicates
cq.where(cb.equals(rootSE.get(SomethingEntity_.somethingId),ID_OF_SOMETHING_ENTITY));

cq.select(rootSE);

entityManager.createQuery(cq).getSingleResult();

---

我知道默认情况下它不起作用。Hibernate中的可选/可为空OneToOne始终是渴望的。是的，也许这是个问题。我自己加载这些关系，然后在不知道我加载了它们的情况下休眠，试图一个接一个地手动加载它们，可能是因为OneToOne很急切。我不希望一个接一个地抓取关系。如果它们被加载并在持久性上下文中，那么Hibernate不会再次加载它们。你的问题需要更多的上下文。你说我不想一个接一个地抓取关系是什么意思？这正是我所拥有的，在Spring数据中，你就是这样做的。你可能需要像这样的实体中的注释@JoinColumn:@OneToOnecascade=CascadeType.ALL，mappedBy=somethingId@JoinColumnname=somethin\u ID private OtherEntity other；