计算德鲁伊中一个维度的不同

Question

我正在尝试区分维度领土的计数。但我认为我得到的与（日期，领土）的计数不同。这是查询：

{
   "queryType": "groupBy",
   "dataSource": "model",
   "granularity": "all",
   "filter": {
   "type": "selector", "dimension": "District", "value": "ATLANTA"},


   "aggregations":[  
      {"name":"!t5", "type":"distinctCount","fieldName":"Territory"},
      {"name":"!T_4","type":"longSum","fieldName":"Units"}
   ],
   "postAggregations":[  
      {  
         "type":"arithmetic",
         "fn":"*",
         "fields":[  
            {  
               "type":"arithmetic",
               "fn":"/",
               "fields":[  
                  {  
                     "type":"fieldAccess",
                     "fieldName":"!T_4"
                  },
                  {  
                     "type":"fieldAccess",
                     "fieldName":"!t5"
                  }
               ]
            },
            {  
               "type":"constant",
               "value":1
            }
         ],
         "name":"AverageDistrict"
      }
   ],
   "intervals": ["2018-06-13T00:00Z/2018-07-25T00:00Z"]
}

所以当我更改日期范围时，distinct(!t5) 的计数也会增加。而在原始数据中，不同的计数在所有日期范围内始终保持不变。我需要的是该日期范围内不同领土值的计数。我希望我已经说清楚了。

Answer 1

尝试使用 'cardinality' 聚合器代替 'distinctCount' 聚合器。希望这有帮助。

计算德鲁伊中一个维度的不同

Count Distinct of a dimension in druid

druid