有可能使用复合键生成spring-data-elasticsearch @Document？

Question

我开始在Spring Boot 1.3.1上使用Spring Data Elasticsearch,我想使用在我的数据库中使用的相同实体,它有复合键.

实体类:

@IdClass(PassengerPk.class)
@Table(name = "passenger")
@Document(indexName="passenger")
public class Passenger implements Serializable {

    @Id
    @ManyToOne
    @JoinColumn(columnDefinition="long", name="user_id", referencedColumnName="id")
    private User user;

    @Id
    @ManyToOne
    @JoinColumn(columnDefinition="long", name="scheduler_id", referencedColumnName="id")
    private Scheduler scheduler;

    @Column(name = "is_active")
    private Boolean isActive;

    ...
}

关键课程:

public class PassengerPk implements Serializable {

    private Long user;
    private Long scheduler;

    public PassengerPk() {
    }

    public PassengerPk(Long user, Long scheduler) {
        this.user = user;
        this.scheduler = scheduler;
    }
    ...
}

JPA Elasticsearch存储库:

public interface PassengerSearchRepository extends ElasticsearchRepository<Passenger, PassengerPk> {

}

数据库: 数据库关系

如果我尝试编译此代码,我会收到此错误.

Caused by: java.lang.IllegalArgumentException: Unsuppored ID type class com.dualion.test.domain.PassengerPk
    at org.springframework.data.elasticsearch.repository.support.ElasticsearchRepositoryFactory.getRepositoryBaseClass(ElasticsearchRepositoryFactory.java:79) ~[spring-data-elasticsearch-1.3.1.RELEASE.jar:na]
    at org.springframework.data.repository.core.support.RepositoryFactorySupport.getRepositoryInformation(RepositoryFactorySupport.java:238) ~[spring-data-commons-1.11.1.RELEASE.jar:na]
    at org.springframework.data.repository.core.support.RepositoryFactorySupport.getRepository(RepositoryFactorySupport.java:181) ~[spring-data-commons-1.11.1.RELEASE.jar:na]
    at org.springframework.data.repository.core.support.RepositoryFactoryBeanSupport.initAndReturn(RepositoryFactoryBeanSupport.java:251) ~[spring-data-commons-1.11.1.RELEASE.jar:na]
    at org.springframework.data.repository.core.support.RepositoryFactoryBeanSupport.afterPropertiesSet(RepositoryFactoryBeanSupport.java:237) ~[spring-data-commons-1.11.1.RELEASE.jar:na]
    at org.springframework.data.elasticsearch.repository.support.ElasticsearchRepositoryFactoryBean.afterPropertiesSet(ElasticsearchRepositoryFactoryBean.java:55) ~[spring-data-elasticsearch-1.3.1.RELEASE.jar:na]
    at org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory.invokeInitMethods(AbstractAutowireCapableBeanFactory.java:1637) ~[spring-beans-4.2.3.RELEASE.jar:4.2.3.RELEASE]
    at org.springframework.beans.factory.support.AbstractAutowireCapableBeanFactory.initializeBean(AbstractAutowireCapableBeanFactory.java:1574) ~[spring-beans-4.2.3.RELEASE.jar:4.2.3.RELEASE]
    ... 71 common frames omitted

我如何修改我的代码？

谢谢

Answer 1

我在其他地方读过相关答案并得出结论这是不可能的；然而，我的固执最终战胜了我，我找到了解决办法。

太长了；强制 spring 使用一个新的存储库，该存储库采用复合 id 的 hashCode 并使用其字符串值作为其 id。

脚步...

创建一个Repository可以处理复合 id 的新对象：

public class HashKeyedRepository<T, ID extends Serializable> extends AbstractElasticsearchRepository<T, ID> {

    public HashKeyedRepository() {
        super();
    }

    public HashKeyedRepository(ElasticsearchEntityInformation<T, ID> metadata,
                                 ElasticsearchOperations elasticsearchOperations) {
        super(metadata, elasticsearchOperations);
    }

    public HashKeyedRepository(ElasticsearchOperations elasticsearchOperations) {
        super(elasticsearchOperations);
    }

    @Override
    protected String stringIdRepresentation(ID id) {
        return String.valueOf(id.hashCode());
    }
}

请注意，这假设您已正确实现.hashCode复合 id 类以正常工作。

接下来，您必须创建一个RepositoryFactoryBean将返回此新值的新值Repository：

public class CustomElasticsearchRepositoryFactoryBean<T extends Repository<S, ID>, S, ID extends Serializable> extends ElasticsearchRepositoryFactoryBean<T, S, ID> {

    private ElasticsearchOperations operations;

    public CustomElasticsearchRepositoryFactoryBean(Class<? extends T> repositoryInterface) {
        super(repositoryInterface);
    }

    public void setElasticsearchOperations(ElasticsearchOperations operations) {
        super.setElasticsearchOperations(operations);
        Assert.notNull(operations);     
        this.operations = operations;
    }

    @Override
    protected RepositoryFactorySupport createRepositoryFactory() {
        return new ElasticsearchRepositoryFactory(operations) {
            @Override
            protected Class<?> getRepositoryBaseClass(RepositoryMetadata metadata) {
                if (!Integer.class.isAssignableFrom(metadata.getIdType()) && !Long.class.isAssignableFrom(metadata.getIdType()) && !Double.class.isAssignableFrom(metadata.getIdType()) && metadata.getIdType() != String.class && metadata.getIdType() != UUID.class) {
                    return HashKeyedRepository.class;
                }
                return super.getRepositoryBaseClass(metadata);
            }
        };
    }
}

最后，在启用存储库时，指定您的新RepositoryFactoryBean类：

@EnableElasticsearchRepositories(basePackages = "xxx.xxx.repository.search", repositoryFactoryBeanClass = CustomElasticsearchRepositoryFactoryBean.class)

如果使用截至本文撰写时支持的任何 ID（即字符串、UUID、数字），则此实现将回退到默认值。我不知道它是否是一个很好的解决方案，因为可能与 .hashCode 发生冲突，但它现在对我有用。

PS我正在使用lomboks@Data为我自动生成 .hashCode 。

PPS我见过其他人（非java）提到的另一个解决方案是对id 的序列化版本（即JSON）进行base64 编码。我认为这可以保证不会发生冲突，但您必须确保删除任何多余的字符（即空格）并保证属性的顺序才能使其有效。