聚合Spring Data Couchbase存储库方法-如何查询给定属性的所有唯一值的列表？

Question

我在弄清楚如何在Spring-Data-Couchbase的存储库中运行一个非常简单的聚合查询时遇到麻烦。我正在使用org.springframework.data.spring-data-couchbase:3.2.0.RELEASE

假设我有一个名为Organism的存储桶，其属性如下：

SpeciesId  SpeciesName  SpeciesDesc     OrderId  OrderDesc  ClassId  ClassDesc
1          Human        H sapiens       21       Primates   31       Mammalia
2          Chimpanzee   P troglodytes   21       Primates   31       Mammalia
3          Fruit fly    D melanogaster  22       Diptera    32       Insecta

我想编写一个查询，使我可以搜索唯一的订单或类；在SQL中，查询看起来像这样：

SELECT OrderId, OrderDesc FROM Organism WHERE OrderDesc like 'Pri%' GROUP BY OrderId, OrderDesc
SELECT ClassId, ClassDesc FROM Organism WHERE ClassDesc like 'Mam%' GROUP BY ClassId, ClassDesc

我将如何在spring-data-couchbase存储库中创建这些查询？

---

到目前为止，我已经尝试了大约一千种不同的方法，但是它们都没有起作用。首先，我尝试向存储库添加不同的查询：

public interface Order {
    public Integer getOrderId();
    public String getOrderDesc();
}

@N1qlPrimaryIndexed
@ViewIndexed(designDoc = "organism")
public interface OrganismRepository extends 
CouchbasePagingAndSortingRepository<Organism, String> {
    Optional<List<Order>> findDistinctByOrderDescStartingWith(String orderDesc);
}

接下来我尝试使用@Query注释：

//Exception - Unable to retrieve enough metadata for N1QL to entity mapping, have you selected _ID and _CAS?
@Query("SELECT orderDesc FROM #{#n1ql.bucket} WHERE orderDesc like 'Pri%' AND #{#n1ql.filter}")
Optional<List<String>> cantSelectPropertyBecauseMissingMetadata();

//Exception - Unable to execute query due to the following n1ql errors:
//   {"msg":"Expression must be a group key or aggregate: (meta(`organism`).`id`)","code":4210}
@Query("#{#n1ql.selectEntity} where #{#n1ql.filter} AND orderDesc like 'pro%' group by orderId, orderDesc")
Optional<List<Order>> selectionsMustBeIncludedInGroupByClause();

//Exception - Unable to retrieve enough metadata for N1QL to entity mapping, have you selected _ID and _CAS?
@Query("SELECT orderId, orderDesc FROM #{#n1ql.bucket} where #{#n1ql.filter} AND orderDesc like 'Pri%' group by orderId, orderDesc")
Optional<List<Order>> cantGroupByJustTwoPropsBecauseMissingMetadata();

//Exception - Unable to retrieve enough metadata for N1QL to entity mapping, have you selected _ID and _CAS?
@Query("SELECT distinct orderDesc FROM #{#n1ql.bucket} WHERE orderDesc like 'pro%' AND #{#n1ql.filter}")
Optional<List<String>> cantSelectDistinctBecauseMissingMetadata();

//Exception - Unable to execute query due to the following n1ql errors:
//  {"msg":"Expression must be a group key or aggregate: (meta(`organism`).`id`)","code":4210}
@Query("SELECT orderId, orderDesc, META(#{#n1ql.bucket}).id as _ID, META(#{#n1ql.bucket}).cas as _CAS FROM #{#n1ql.bucket} where #{#n1ql.filter} AND orderDesc like 'Pri%' group by orderId, orderDesc")
Optional<List<Order>> cantGroupByJustTwoPropsBecauseStillMissingMetadata();

//Exception - Unable to execute query due to the following n1ql errors:
//  {"msg":"Expression must be a group key or aggregate: (meta(`organism`).`id`)","code":4210}
@Query("SELECT orderId, orderDesc, META(#{#n1ql.bucket}).id as _ID, META(#{#n1ql.bucket}).cas as _CAS FROM #{#n1ql.bucket} where #{#n1ql.filter} AND orderDesc like 'Pri%' group by orderId, orderDesc, _ID, _CAS")
Optional<List<Order>> cantGroupByMETA();

@Query("SELECT distinct orderDesc, META(#{#n1ql.bucket}).id as _ID, META(#{#n1ql.bucket}).cas as _CAS FROM #{#n1ql.bucket} WHERE orderDesc like 'Pri%' AND #{#n1ql.filter}")
Optional<List<Order>> resultsAreNotDistinct();

@Query("SELECT distinct orderId, orderDesc, META(#{#n1ql.bucket}).id as _ID, META(#{#n1ql.bucket}).cas as _CAS FROM #{#n1ql.bucket} WHERE orderDesc like 'Pri%' AND #{#n1ql.filter}")
Optional<List<Order>> resultsAreNotDistinct2();

@Query("SELECT orderId, orderDesc, META(#{#n1ql.bucket}).id as _ID, META(#{#n1ql.bucket}).cas as _CAS FROM #{#n1ql.bucket} where #{#n1ql.filter} AND orderDesc like 'Pri%' group by orderId, orderDesc, META(#{#n1ql.bucket}).id, META(#{#n1ql.bucket}).cas")
Optional<List<Order>> resultsAreNotDistinct3();

---

到目前为止，使此功能正常工作的唯一方法是使用com.couchbase.client.java.Bucket中的com.couchbase.client.java-client:2.7.9。但是如果Spring-Data-Couchbase中没有办法做到这一点，我会感到震惊...

String queryString = "SELECT distinct orderId, orderDesc " + 
    "FROM `organisms` " + 
    "WHERE orderDesc LIKE $search"
Cluster cluster = CouchbaseCluster.create(hosts).authenticate(bucketUser, bucketPassword);
Bucket bucket = cluster.openBucket(bucketName);
JsonObject parameters = JsonObject.create().put("search", search);
ParameterizedN1qlQuery q = N1qlQuery.parameterized(queryString, parameters);
N1qlQueryResult result = bucket.query(q);

Answer 1

[Couchbase Spring Data的当前实现将始终尝试获取CAS和文档ID（因为元数据对于某些用例而言很重要），这就是您收到此错误的原因。

如果只返回所需的属性，我建议使用类似于以下内容的语法：

@Override
public List<KeyValueVO> listKeyValue(String companyId) {
    String bucketName = securityGroupRepository.getCouchbaseOperations().getCouchbaseBucket().name();
    String queryString = "SELECT meta().id as id, name FROM "+bucketName+
            " WHERE  _class = '"+SecurityGroup.class.getName() +"' and removed = false and companyId = '"+companyId+"' order by name ";
    N1qlParams params = N1qlParams.build().consistency(ScanConsistency.NOT_BOUNDED).adhoc(true);
    ParameterizedN1qlQuery query = N1qlQuery.parameterized(queryString, JsonObject.create(), params);
    return securityGroupRepository.getCouchbaseOperations().findByN1QLProjection(query, KeyValueVO.class);

}

PS：是的，我应该正确使用ParameterizedN1qlQuery而不是在字符串中连接参数，但是您明白了。