基于嵌套属性的数组属性的文档的唯一计数

Question

假设它是下面我的索引的映射，我想创建一个弹性聚合查询来计算

test

数组属性

的每个元素的文档数量

• 映射

"mappings": {
      "properties": {
        "code": {
          "type": "keyword"
        },
        "details": {
          "type": "nested",
          "properties": {
            "name": {
              "type": "text",
              "analyzer": "html_stripper"
            },            
            "elts": {
              "type": "integer"
            }
          }
        }
      }
    },

• 文档示例

[
    {
        "code": "test",
        "details": [
            {"name": "test", "elts": [1,2,3,4]}
        ]
    },
    {
        "code": "test2",
        "details": [
            {"name": "test2", "elts": [1,2,4]}
        ]
    }
]

我已经尝试过这个查询

POST /my-index/_search

{
  "size": 0,
  "aggs": {
    "unique_count": {
      "terms": {
        "field": "details.elts"
      }
    }
  }
}

但是它不起作用。下面是结果

{
  "took": 2,
  "timed_out": false,
  "_shards": {
    "total": 1,
    "successful": 1,
    "skipped": 0,
    "failed": 0
  },
  "hits": {
    "total": {
      "value": 3,
      "relation": "eq"
    },
    "max_score": null,
    "hits": []
  },
  "aggregations": {
    "lterms#unique_count": {
      "doc_count_error_upper_bound": 0,
      "sum_other_doc_count": 0,
      "buckets": []
    }
  }
}

我期待的是一个像水桶一样的东西

[
  {
    "key": "1",
    "doc_count": 2
  },
  {
    "key": "2",
    "doc_count": 2
  },
  {
    "key": "3",
    "doc_count": 1
  },
  {
    "key": "4",
    "doc_count": 2
  }
]

Answer 1

您应该通过

nested

聚合来包装聚合以访问嵌套文档

POST /my-index/_search?filter_path=aggregations
{
  "aggs": {
    "inside_details": {
      "nested": {
        "path": "details"
      },
      "aggs": {
        "unique_count": {
          "terms": {
            "field": "details.elts",
            "order": {
              "_key": "asc"
            }
          }
        }
      }
    }
  }
}

回应

{
  "aggregations" : {
    "inside_details" : {
      "doc_count" : 2,
      "unique_count" : {
        "doc_count_error_upper_bound" : 0,
        "sum_other_doc_count" : 0,
        "buckets" : [
          {
            "key" : 1,
            "doc_count" : 2
          },
          {
            "key" : 2,
            "doc_count" : 2
          },
          {
            "key" : 3,
            "doc_count" : 1
          },
          {
            "key" : 4,
            "doc_count" : 2
          }
        ]
      }
    }
  }
}