【es】cardinality 计算不准确问题

遇到问题：

两个结果不一样，按说是一样的

结果一：

    
    
{
  "query": {
    "bool": {
      "must_not": [
        {
          "match_phrase": {
            "reqUA": "Jakarta Commons-HttpClient/3.1"
          }
        },
        {
          "match_phrase": {
            "reqReferer": "http://www.baidu.com/s?wd=www"
          }
        }
      ],
      "must": [
        {
          "range": {
            "reqTime": {
              "gte": "2016-09-25 22:00:00",
              "lte": "2016-09-26 22:00:00"
            }
          }
        },
        {
          "range": {
            "operateBeforeObj.sendTime": {
              "gte": "2016-09-25 22:00:00",
              "lte": "2016-09-26 22:00:00"
            }
          }
        },
        {
          "terms": {
            "productPageCode": [
              "10001",
              "33002"
            ]
          }
        }
      ]
    }
  },
  "from": 0,
  "aggs": {
    "channelTag": {
      "terms": {
        "field": "channelTag",
        "size": 0
      },
      "aggs": {
        "userId": {
          "cardinality": {
            "field": "user.userId"
          }
        }
      }
    }
  },
  "size": 0
}

结果二：

    
    
{
  "query": {
    "bool": {
      "must_not": [
        {
          "match_phrase": {
            "reqUA": "Jakarta Commons-HttpClient/3.1"
          }
        },
        {
          "match_phrase": {
            "reqReferer": "http://www.baidu.com/s?wd=www"
          }
        }
      ],
      "must": [
        {
          "range": {
            "reqTime": {
              "gte": "2016-09-25 22:00:00",
              "lte": "2016-09-26 22:00:00"
            }
          }
        },
        {
          "range": {
            "operateBeforeObj.sendTime": {
              "gte": "2016-09-25 22:00:00",
              "lte": "2016-09-26 22:00:00"
            }
          }
        },
        {
          "terms": {
            "productPageCode": [
              "10001",
              "33002"
            ]
          }
        }
      ]
    }
  },
  "from": 0,
  "aggs": {
    "userId": {
      "cardinality": {
        "field": "user.userId"
      }
    }
  },
  "size": 0
}

分析问题：

问题应该在cardinality上，cardinality有个参数 "precision_threshold": 100 ，100是个预设值，你的真实值小于100计算出来的值就是正确的，真实值大于100计算出来的值就是模糊的，100可以自定义。

解决问题：

   
   
{
    "aggs":{
        "author_count":{
            "cardinality":{
                "field":"author_hash",
                "precision_threshold":100
            }
        }
    }
}

参考：

https://www.elastic.co/guide/en/elasticsearch/reference/current/search-aggregations-metrics-cardinality-aggregation.html