Hibernate Search 5.5.5.Final：如何基于关联字段排序模糊搜索结果

Got it, let's tackle this sorting issue with Hibernate Search 5.5.5.Final. Since you already have the fuzzy search working, adding sorting by fleetCustomer.customerId just requires a few tweaks to your entity mappings and query setup.

1. 完善实体的索引映射 #

First, we need to make sure the customerId from the associated fleetCustomer is properly embedded into the Vehicle's index. Here's how to adjust your entities:

VehicleDao 实体调整 #

Add the @IndexedEmbedded annotation to your fleetCustomer association, specifying the field we need for sorting:

@Entity
@Table(name = "VEHICLE")
@Indexed
public class VehicleDao {
    // 已有的字段...
    @Id
    @GeneratedValue(strategy = GenerationType.AUTO)
    @Column(name = "vehicle_id")
    private Long vehicleId;

    @Field // 确保licensePlate被索引，支持模糊搜索
    @Column(name = "license_plate")
    private String licensePlate;

    // 关键配置：将fleetCustomer的customerId嵌入到当前实体的索引中
    @ManyToOne
    @IndexedEmbedded(includePaths = "customerId")
    private FleetCustomerDao fleetCustomer;

    // 其他字段、getter和setter...
}

FleetCustomerDao 实体调整 #

Make sure the customerId field is marked with @Field so it can be indexed and used for sorting:

@Entity
@Table(name = "FLEET_CUSTOMER")
public class FleetCustomerDao {
    @Id
    @Column(name = "customer_id")
    @Field // 必须添加此注解，让customerId可被索引和排序
    private Long customerId;

    // 其他字段、getter和setter...
}

2. 在搜索查询中添加排序逻辑 #

Now that the indexing is set up, you just need to add the sorting rule to your existing fuzzy search query:

// 获取FullTextEntityManager实例
FullTextEntityManager fullTextEntityManager = Search.getFullTextEntityManager(entityManager);

// 构建模糊搜索的Lucene查询（你已有的模糊搜索逻辑，这里做示例参考）
QueryBuilder queryBuilder = fullTextEntityManager.getSearchFactory()
        .buildQueryBuilder()
        .forEntity(VehicleDao.class)
        .get();

// 模糊搜索licensePlate的逻辑
org.apache.lucene.search.Query fuzzyQuery = queryBuilder
        .keyword()
        .fuzzy()
        .onField("licensePlate")
        .matching("your-search-term") // 替换为实际的用户搜索关键词
        .createQuery();

// 创建排序规则：按fleetCustomer.customerId升序排列（false=升序，true=降序）
Sort sort = new Sort(
        new SortField("fleetCustomer.customerId", SortField.Type.LONG, false)
);

// 构建FullTextQuery并绑定排序规则
FullTextQuery fullTextQuery = fullTextEntityManager.createFullTextQuery(fuzzyQuery, VehicleDao.class);
fullTextQuery.setSort(sort); // 关键步骤：给查询添加排序

// 执行查询，得到排序后的结果
List<VehicleDao> sortedResults = fullTextQuery.getResultList();

关键注意事项 #

• 字段类型匹配：如果你的customerId是字符串类型，把SortField.Type.LONG改成SortField.Type.STRING。
• 索引更新：当FleetCustomer的customerId变化时，Hibernate Search会自动更新关联Vehicle的索引（默认行为），但如果用了延迟加载，要确保关联实体被正确加载，或者配置@IndexedEmbedded的fetch属性。
• 多字段排序：如果需要同时按多个字段排序，只需在Sort的构造参数中添加多个SortField，比如：

  Sort sort = new Sort(
          new SortField("fleetCustomer.customerId", SortField.Type.LONG, false),
          new SortField("licensePlate", SortField.Type.STRING, true)
  );
  

内容的提问来源于stack exchange，提问作者aurelius