匹配对象数组中多个字段匹配的文档

Question

我在 ES 的文档中有一个对象数组

我只想匹配对象（数组内）内的 2 个字段匹配的文档。

我的问题是这个。假设我有 3 个文档（删除了外部文档），其中有一组对象，例如

                       // Doc 1
                        "questionAnswers": [
               
                        {
                            "question": "First yes no question",
                            "answer": "Yes"
                        },
                        {
                            "question": "Second yes no question",
                            "answer": "No"
                        }
                    ]


                       // Doc 2
                        "questionAnswers": [
               
                        {
                            "question": "First yes no question",
                            "answer": "No"
                        },
                        {
                            "question": "Second yes no question",
                            "answer": "No"
                        }
                    ]


                       // Doc 3
                        "questionAnswers": [
               
                        {
                            "question": "First yes no question",
                            "answer": "No"
                        },
                        {
                            "question": "Second yes no question",
                            "answer": "Yes"
                        }
                    ]

我的查询是

{
"from": 0,
"size": 25,
"query": {
    "bool": {
        "filter": [
            {
                "match": {
                    "questionAnswers.question.keyword": "First yes no question"
                }
            },
            {
                    "match": {
                        "questionAnswers.answer": "Yes"
                    }
                
            }
        ]
    }
}

}

我目前正在获取文档 1 和 3 的匹配项。而我只希望第一个文档匹配 where question = First yes no question and answer = Yes

如何做到这一点？

Answer 1

您需要将questionAnswers字段定义为nested type，以便可以单独查询其中的每个数组对象，并且可以使用inner_hits在结果中仅获取匹配的文档。

添加一个带有索引映射、搜索查询和搜索结果的工作示例

索引映射：

{
  "mappings": {
    "properties": {
      "questionAnswers": {
        "type": "nested"
      }
    }
  }
}

搜索查询：

{
  "query": {
    "nested": {
      "path": "questionAnswers",
      "query": {
        "bool": {
          "filter": [
            {
              "match": {
                "questionAnswers.question.keyword": "First yes no question"
              }
            },
            {
              "match": {
                "questionAnswers.answer": "Yes"
              }
            }
          ]
        }
      },
      "score_mode": "avg",
      "inner_hits": {}
    }
  }
}

搜索结果：

"hits": [
      {
        "_index": "68760699",
        "_type": "_doc",
        "_id": "1",
        "_score": 0.0,
        "_source": {
          "questionAnswers": [
            {
              "question": "First yes no question",
              "answer": "Yes"
            },
            {
              "question": "Second yes no question",
              "answer": "No"
            }
          ]
        },
        "inner_hits": {
          "questionAnswers": {
            "hits": {
              "total": {
                "value": 1,
                "relation": "eq"
              },
              "max_score": 0.0,
              "hits": [
                {
                  "_index": "68760699",
                  "_type": "_doc",
                  "_id": "1",
                  "_nested": {
                    "field": "questionAnswers",
                    "offset": 0
                  },
                  "_score": 0.0,
                  "_source": {
                    "question": "First yes no question",           //note this
                    "answer": "Yes"
                  }
                }
              ]
            }
          }
        }
      }
    ]