多级嵌套查询

Question

我有一个多层嵌套文档。我想基于多个嵌套查询进行查询，所有条件都必须匹配。

例子

文档 1

{
    "publishing_rule": {
        "publishing_orders": [{
                "transporters": [{
                    "fteid": "81"
                }],
                "active": false
            },
            {
                "transporters": [{
                    "fteid": "82"
                }],
                "active": true
            }
        ]
    }
}

文档 2

{
    "publishing_rule": {
        "publishing_orders": [{
                "transporters": [{
                    "fteid": "81"
                }],
                "active": true
            },
            {
                "transporters": [{
                    "fteid": "82"
                }],
                "active": false
            }
        ]
    }
}

我想获取符合以下条件的所有文档

publishing_rule.publishing_orders.active = true 

AND 

publishing_rule.publishing_orders.transporters.fteid = '81'

active 和 transporters.fteid 应该是同一对象的一部分。

我尝试创建以下映射

{
  "mappings": {
    "_doc": {
      "properties": {
        "publishing_rule.publishing_orders": {
          "type": "nested",
          "properties": {
            "transporters": {
              "type": "nested"
            }
          }
        }
      }
    }
  }
}

并在下面的查询中使用

{
  "query": {
    "bool": {
      "must": [
        {
          "nested": {
            "path": "publishing_rule.publishing_orders",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "publishing_rule.publishing_orders.active": true
                    }
                  }
                ]
              }
            }
          }
        },
        {
          "nested": {
            "path": "publishing_rule.publishing_orders.transporters",
            "query": {
              "bool": {
                "must": [
                  {
                    "match": {
                      "publishing_rule.publishing_orders.transporters.fteid": "81"
                    }
                  }
                ]
              }
            }
          }
        }
      ]
    }
  }
}

但我没有得到预期的结果。返回两个文档的查询。

我希望结果中只有文档 2。

Answer 1

您的查询实际上会查找匹配 active = true 或 fteid = 81 但不会同时匹配两者的任何文档。文档 1 和文档 2 满足了这些条件。这就是您获得这两个的原因。

这个查询应该有效

{
  "query": {
    "nested": {
      "path": "publishing_rule.publishing_orders",
      "query": {
        "bool": {
          "must": [
            {
              "match": {
                "publishing_rule.publishing_orders.active": true
              }
            },
            {
              "nested": {
                "path": "publishing_rule.publishing_orders.transporters",
                "query": {
                  "bool": {
                    "must": [
                      {
                        "match": {
                          "publishing_rule.publishing_orders.transporters.fteid": "81"
                        }
                      }
                    ]
                  }
                }
              }
            }
          ]
        }
      }
    }
  }
}

请注意，我使用单个 nested 作为入口点，然后使用内部 nested，这是为了让两个过滤器一起在文档中进行搜索。

更新

映射

GET /myindex/_mapping

{
    "mappings": {
      "_doc": {
        "properties": {
          "publishing_rule": {
            "properties": {
              "publishing_orders": {
                "type": "nested",
                "properties": {
                  "active": {
                    "type": "boolean"
                  },
                  "transporters": {
                    "type": "nested",
                    "properties": {
                      "fteid": {
                        "type": "text",
                        "fields": {
                          "keyword": {
                            "type": "keyword",
                            "ignore_above": 256
                          }
                        }
                      }
                    }
                  }
                }
              }
            }
          }
        }
      }
    }

ES 结果

文件 2

{
  "took": 6,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "failed": 0
  },
  "hits": {
    "total": 1,
    "max_score": 1.89712,
    "hits": [
      {
        "_index": "myindex",
        "_type": "_doc",
        "_id": "AWzu__bgqsCjtPMt7kG_",
        "_score": 1.89712,
        "_source": {
          "publishing_rule": {
            "publishing_orders": [
              {
                "transporters": [
                  {
                    "fteid": "81" // matched
                  }
                ],
                "active": true // matched
              },
              {
                "transporters": [
                  {
                    "fteid": "82"
                  }
                ],
                "active": true
              }
            ]
          }
        }
      }
    ]
  }
}

希望对您有所帮助

Answer 2

我看到@deerawan 的回答足以解决您提出的问题。由于您不接受他的回答，我相信您正在寻找的是在结果中单独获取嵌套文档。我修改了@deerawan 的查询以单独包含与查询匹配的嵌套文档

{
  "_source": false,
  "query": {
    "nested": {
      "path": "publishing_rule.publishing_orders",
      "query": {
        "bool": {
          "must": [
            {
              "match": {
                "publishing_rule.publishing_orders.active": true
              }
            },
            {
              "nested": {
                "path": "publishing_rule.publishing_orders.transporters",
                "query": {
                  "bool": {
                    "must": [
                      {
                        "match": {
                          "publishing_rule.publishing_orders.transporters.fteid": "81"
                        }
                      }
                    ]
                  }
                }
              }
            }
          ]
        }
      },
      "inner_hits": {}
    }
  }
}

这应该会给您以下响应

{
  "took": 7,
  "timed_out": false,
  "_shards": {
    "total": 5,
    "successful": 5,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": 1,
    "max_score": 1.3862944,
    "hits": [
      {
        "_index": "nest",
        "_type": "doc",
        "_id": "1234",
        "_score": 1.3862944,
        "inner_hits": {
          "publishing_rule.publishing_orders": {
            "hits": {
              "total": 1,
              "max_score": 1.3862944,
              "hits": [
                {
                  "_nested": {
                    "field": "publishing_rule.publishing_orders",
                    "offset": 0
                  },
                  "_score": 1.3862944,
                  "_source": {
                    "transporters": [
                      {
                        "fteid": "81"
                      }
                    ],
                    "active": true
                  }
                }
              ]
            }
          }
        }
      }
    ]
  }
}

多级嵌套查询

Multi Level Nested Queries

elasticsearch

kibana