关于Druid中的聚合后度量的TopN查询

Question

我需要对后汇总指标进行前n个查询。听起来很简单。这是我正在使用的查询-

{
     "queryType": "topN",
     "dataSource": "model",
     "granularity": "all",
     "intervals": [
       "2018-01-01T00:00:00/2018-12-31T23:59:59"
     ],
     "dimension": {
       "type": "lookup",
       "dimension": "COName",
       "outputName": "COName",
       "name": "COName"
     },
    "metric": "MarketShare",

     "threshold": 10,
     "aggregations": [
       {
         "type": "filtered",
         "filter": {
           "type": "selector",
           "dimension": "Brand",
           "value": "Nike"
         },
         "aggregator": {
           "type": "doubleSum",
           "name": "nike_sum",
           "fieldName": "Sales"
         }
       },
       {
         "type": "doubleSum",
         "name": "total_sum",
         "fieldName": "Sales"
       }
     ],
      "postAggregations": [
        {
          "type": "arithmetic",
          "name": "MarketShare",
          "fn": "*",
          "fields": [
            {
              "type": "arithmetic",
              "name": "div",
              "fn": "/",
              "fields": [
                {
                  "type": "fieldAccess",
                  "name": "brand_sum",
                  "fieldName": "nike_sum"
                },
                {
                  "type": "fieldAccess",
                  "name": "total_sum_1",
                  "fieldName": "total_sum"
                }
              ]
            },
            {
              "type": "constant",
              "name": "const",
              "value": 100
            }
          ]
        }
      ],
     "context": {
       "skipEmptyBuckets": true
     }
}

这给了我奇怪的前N个结果（数字不正确）。我敢肯定，因为当我对某些特定的COName进行简单分组时，我得到了预期的结果。 有趣的事实：
当我将"metric": "MarketShare"替换为诸如"metric": "total_sum"之类的其他度量标准时，我得到的数字正确，但是排序现在使用“ total_sum”。