如何使用Spring Data删除MongoDB文档数组内匹配条件的元素

需要在满足指定过滤条件时，删除MongoDB文档内caseDetails数组的匹配子元素：现有数千条记录，每条均包含caseDetails数组，数组内存有多个子项，传入caseId参数后，删除所有符合过滤条件的文档中caseDetails数组里caseId与传入值匹配的子项，无需额外编写新查询逻辑。

初始数据库示例数据 #

{
    "_id": "60f8fcbe8f67d30784c45561",
    "messageType": "e",
    "sendDate": "2021-07-22T05:06:04.023Z",
    "assigneeId": 600003,
    "customerId": 58004,
    "subject": "Email Test 1[#24894006]",
    "mailAnswered": false,
    "hasAttachments": false,
    "caseDetails": [
        {
            "caseId": 24894006,
            "employeeId": 600003
        },
        {
            "caseId": 34894006,
            "employeeId": 600003
        },
        {
            "caseId": 44894006,
            "employeeId": 600003
        }

    ]
},  
{
    "_id": "60f8fcbe8f67d30784c45561",
    "messageType": "e",
    "sendDate": "2021-07-22T05:06:04.023Z",
    "assigneeId": 600003,
    "customerId": 58004,
    "subject": "Email Test 1[#24894006]",
    "mailAnswered": false,
    "hasAttachments": false,
    "caseDetails": [
        {
            "caseId": 24894006,
            "employeeId": 600003
        },
        {
            "caseId": 34894006,
            "employeeId": 600003
        },
        {
            "caseId": 44894006,
            "employeeId": 600003
        }
    ]
}

预期结果（删除caseId=24894006的对应子项后） #

{
    "_id": "60f8fcbe8f67d30784c45561",
    "messageType": "e",
    "sendDate": "2021-07-22T05:06:04.023Z",
    "assigneeId": 600003,
    "customerId": 58004,
    "subject": "Email Test 1[#24894006]",
    "mailAnswered": false,
    "hasAttachments": false,
    "caseDetails": [
        {
            "caseId": 34894006,
            "employeeId": 600003
        },
        {
            "caseId": 44894006,
            "employeeId": 600003
        }

    ]
},  
{
    "_id": "80f8fcbe8f67d30784c45561",
    "messageType": "e",
    "sendDate": "2021-07-22T05:06:04.023Z",
    "assigneeId": 600003,
    "customerId": 58004,
    "subject": "Email Test 1[#24894006]",
    "mailAnswered": false,
    "hasAttachments": false,
    "caseDetails": [
        {
            "caseId": 54894006,
            "employeeId": 600003
        },
        {
            "caseId": 64894006,
            "employeeId": 600003
        }
    ]
}

现有代码 #

接口实现代码 #

@Override
  public List<Email> unlinkMessageCaseDetailsByCaseId(MessageFilter messageFilter) {
    Criteria finalCriteria = handleCriteria(messageFilter);
    Query query = query(finalCriteria);
    // Update update = new Update().pull("caseDetails", Collections.singletonMap("caseId", messageFilter.getCaseId()));
    List<Email> uMono = mongoTemplate.find(query, Email.class).collectList().block();
    return uMono;
  }

Repository定义 #

@MongoRepository
public interface EmailRepository 
         extends ReactiveCrudRepository<Email, String>,
                 MessageRepositoryCustom {

}

你注释中的pull操作思路是正确的，只需要补充执行更新的逻辑即可，完整实现如下：

@Override
public List<Email> unlinkMessageCaseDetailsByCaseId(MessageFilter messageFilter) {
    Criteria finalCriteria = handleCriteria(messageFilter);
    Query query = query(finalCriteria);
    // 构造数组删除规则：删除caseDetails中caseId匹配传入参数的子项
    Update update = new Update().pull("caseDetails", Collections.singletonMap("caseId", messageFilter.getCaseId()));
    // 批量更新所有符合条件的文档
    mongoTemplate.updateMulti(query, update, Email.class).block();
    // 返回更新后的匹配文档列表
    return mongoTemplate.find(query, Email.class).collectList().block();
}

补充说明 #

• pull是MongoDB原生的数组操作符，会自动遍历所有匹配文档的caseDetails数组，删除所有符合条件的子元素，不需要手动遍历文档，性能足以应对数千条量级的数据
• 如果需要返回更新前的文档，可以使用findAndModify方法，配置ReturnDocument.BEFORE参数即可将更新和查询合并为一次请求
• 如果不需要同步返回结果，建议移除block()调用直接返回Flux<Email>，符合响应式编程规范

内容的提问来源于stack exchange，提问作者Shrey Soni