我正在尝试从Datastudio内部准备的XLS复制数据,
维度为YYYYMM,输出为MonthlyRate。 基本上,累积响应/累积会议=月费率
我应该如何在Datastudio中准备计算出的字段和图形?
我当前使用的计算字段是Sum(Response)/ Count(Meetings),使用维度YYYYMM(YearMonth)。 场上的移动平均值,但是最终导致数字稍微偏斜。 例如201809的0.45变为0.47。
答案 0 :(得分:2)
以下示例适用于BigQuery标准SQL
#standardSQL
WITH `project.dataset.table` AS (
SELECT '201712' yyyymm, 4580 SumResponse, 6741 CountMeetings UNION ALL
SELECT '201801', 3574, 6926 UNION ALL
SELECT '201802', 2020, 6433 UNION ALL
SELECT '201803', 1895, 6635 UNION ALL
SELECT '201804', 2174, 6163 UNION ALL
SELECT '201805', 3058, 7697 UNION ALL
SELECT '201806', 3313, 7838 UNION ALL
SELECT '201807', 4043, 8586 UNION ALL
SELECT '201808', 5053, 9355 UNION ALL
SELECT '201809', 1122, 1300
)
SELECT
yyyymm,
SumResponse,
SUM(SumResponse) OVER(ORDER BY yyyymm) CumulativeResponse ,
CountMeetings,
SUM(CountMeetings) OVER(ORDER BY yyyymm) CumulativeMeetings,
SUM(SumResponse) OVER(ORDER BY yyyymm)/SUM(CountMeetings) OVER(ORDER BY yyyymm) MonthlyRate
FROM `project.dataset.table`
ORDER BY yyyymm
结果:
Row yyyymm SumResponse CumulativeResponse CountMeetings CumulativeMeetings MonthlyRate
1 201712 4580 4580 6741 6741 0.6794244177421748
2 201801 3574 8154 6926 13667 0.596619594644033
3 201802 2020 10174 6433 20100 0.5061691542288557
4 201803 1895 12069 6635 26735 0.45143070880867775
5 201804 2174 14243 6163 32898 0.4329442519302085
6 201805 3058 17301 7697 40595 0.4261854908239931
7 201806 3313 20614 7838 48433 0.4256188962071315
8 201807 4043 24657 8586 57019 0.4324348024342763
9 201808 5053 29710 9355 66374 0.4476150299816193
10 201809 1122 30832 1300 67674 0.4555959452670154