Spring Boot中MongoDB对象数组查询报$elemMatch错误求助

问题背景 #

现有指定结构的MongoDB集合文档，需实现短语搜索+检查meta数组中存在指定index_field_id的查询功能，但编写的Spring Data MongoDB查询方法运行时抛出异常。

MongoDB文档结构 #

{
  "_id": "ba4d5a0e-1759-4e88-bd57-5e7781d46da7",
  "name": "Doc 3 updated",
  "documentClass": "033a1f67-b382-43f5-b1e4-7ade26f1522c",
  "documentClassName": "Sample Document",
  "meta": [
    {
      "index_field_id": "9e744f36-1c63-45e8-8180-3ceef894b724",
      "value": "a"
    },
    {
      "index_field_id": "08afba06-9a7e-4df8-a946-a8cccbefda8b",
      "value": "asafsf"
    }
  ],
  "version": 2,
  "archived": false,
  "checkedOut": false,
  "checkedOutBy": "svpadm",
  "created": {
    "$date": "2023-12-24T11:12:39.580Z"
  },
  "modified": {
    "$date": "2024-01-09T19:48:29.785Z"
  },
  "owner": "79866869-398a-4920-975b-032decc12461",
  "ownerClear": "svpadm",
  "tenant": "ECMDEV01-T",
  "updatedBy": "79866869-398a-4920-975b-032decc12461",
  "updatedByClear": "svpadm",
  "_class": "de.metasonic.cloud.models.database.Document"
}

需求说明 #

• 通过指定短语搜索文档，匹配字段包括：name、documentClass、documentClassName、workbookId、owner、ownerClear、tenant、updatedBy、updatedByClear、linkedDocuments、providerState、providerMetaData，匹配时忽略大小写。
• 同时筛选出meta数组中存在indexFieldIds列表中任一index_field_id的文档。

原查询方法 #

@Query("{$and: ["
            + " {$or: ["
            + "  {'name': {$regex: ?0, $options: 'i'}}, "
            + "  {'documentClass': {$regex: ?0, $options: 'i'}}, "
            + "  {'documentClassName': {$regex: ?0, $options: 'i'}}, "
            + "  {'workbookId': {$regex: ?0, $options: 'i'}}, "
            + "  {'owner': {$regex: ?0, $options: 'i'}}, "
            + "  {'ownerClear': {$regex: ?0, $options: 'i'}}, "
            + "  {'tenant': {$regex: ?0, $options: 'i'}}, "
            + "  {'updatedBy': {$regex: ?0, $options: 'i'}}, "
            + "  {'updatedByClear': {$regex: ?0, $options: 'i'}}, "
            + "  {'linkedDocuments': {$elemMatch: {$regex: ?0, $options: 'i'}}}, "
            + "  {'providerState': {$regex: ?0, $options: 'i'}}, "
            + "  {'providerMetaData': {$elemMatch: {$regex: ?0, $options: 'i'}}}"
            + " ]},
"            + " {'meta': {$elemMatch: {'index_field_id': {$in: ?1}}}}"
            + "]}")
Page<Document> searchByPhraseAndIndexFieldIds(String searchPhrase, List<String> indexFieldIds, Pageable pageable);

对应的DTO类 #

public class DocumentMetaDTO {
    String index_field_id;
    Object value;
    ArrayList<DocumentMetaDTO> categoryValues;
}

异常信息 #

Exception : org.springframework.data.mongodb.UncategorizedMongoDbException: Query failed with error code 2 with name 'BadValue' and error message '$elemMatch needs an Object' on server localhost:27018; nested exception is com.mongodb.MongoQueryException: Query failed with error code 2 with name 'BadValue' and error message '$elemMatch needs an Object' on server localhost:27018

问题原因 #

报错核心是错误使用$elemMatch操作符：

1. $elemMatch用于匹配数组中嵌套对象同时满足多个条件的元素，若只是匹配数组中的字符串元素是否符合正则，无需使用$elemMatch，直接对数组字段用正则即可，MongoDB会自动匹配数组中任意满足条件的元素。原查询中linkedDocuments和providerMetaData字段若为字符串数组，用$elemMatch包裹正则属于错误用法。
2. 针对meta数组的查询仅需匹配index_field_id在指定列表中，无需$elemMatch，直接通过meta.index_field_id路径查询更简洁高效。

修正后的查询方法 #

@Query("{$and: ["
            + " {$or: ["
            + "  {'name': {$regex: ?0, $options: 'i'}}, "
            + "  {'documentClass': {$regex: ?0, $options: 'i'}}, "
            + "  {'documentClassName': {$regex: ?0, $options: 'i'}}, "
            + "  {'workbookId': {$regex: ?0, $options: 'i'}}, "
            + "  {'owner': {$regex: ?0, $options: 'i'}}, "
            + "  {'ownerClear': {$regex: ?0, $options: 'i'}}, "
            + "  {'tenant': {$regex: ?0, $options: 'i'}}, "
            + "  {'updatedBy': {$regex: ?0, $options: 'i'}}, "
            + "  {'updatedByClear': {$regex: ?0, $options: 'i'}}, "
            + "  {'linkedDocuments': {$regex: ?0, $options: 'i'}}, "
            + "  {'providerState': {$regex: ?0, $options: 'i'}}, "
            + "  {'providerMetaData': {$regex: ?0, $options: 'i'}}"
            + " ]},
"            + " {'meta.index_field_id': {$in: ?1}}"
            + "]}")
Page<Document> searchByPhraseAndIndexFieldIds(String searchPhrase, List<String> indexFieldIds, Pageable pageable);

修正点说明 #

• 移除linkedDocuments和providerMetaData字段查询中的$elemMatch，直接使用正则匹配数组元素。
• 将meta数组的查询从{'meta': {$elemMatch: {'index_field_id': {$in: ?1}}}}改为{'meta.index_field_id': {$in: ?1}}，无需$elemMatch即可实现匹配数组中任一元素的index_field_id在指定列表中的需求。

内容的提问来源于stack exchange，提问作者Viranga