根据其他表中的份额拆分价值

Question

DB-小提琴

CREATE TABLE costs (
    id SERIAL PRIMARY KEY,
    event_date DATE,
    country VARCHAR,
    channel VARCHAR,
    costs_per_day_per_country DECIMAL);

INSERT INTO costs
(event_date, country, channel, costs_per_day_per_country)
VALUES 
('2020-02-08', 'DE', 'channel_01', '400'),
('2020-02-08', 'DE', 'channel_02', '400'),
('2020-02-08', 'FR', 'channel_01', '400'),
('2020-02-08', 'FR', 'channel_02', '400'),
('2020-02-08', 'NL', 'channel_01', '400'),
('2020-02-08', 'NL', 'channel_02', '400'),

('2020-04-15', 'DE', 'channel_01', '300'),
('2020-04-15', 'DE', 'channel_02', '300'),
('2020-04-15', 'FR', 'channel_01', '300'),
('2020-04-15', 'FR', 'channel_02', '300'),
('2020-04-15', 'NL', 'channel_01', '300'),
('2020-04-15', 'NL', 'channel_02', '300');


CREATE TABLE revenue (
    id SERIAL PRIMARY KEY,
    event_date DATE,
    country VARCHAR,
    revenue_share DECIMAL);

INSERT INTO revenue
(event_date, country, revenue_share)
VALUES 
('2020-02-08', 'DE', '0.60'),
('2020-02-08', 'FR', '0.30'),
('2020-02-08', 'NL', '0.10'),

('2020-04-15', 'DE', '0.60'),
('2020-04-15', 'FR', '0.30'),
('2020-04-15', 'NL', '0.10');

预期结果：

event_date    |    country   |     channel       |            costs              |
--------------|--------------|-------------------|-------------------------------|---------
2020-02-08    |      DE      |     channel_01    |      120  (=400 x 0.6 / 2)    | 
2020-02-08    |      DE      |     channel_02    |      120  (=400 x 0.6 / 2)    |   
2020-02-08    |      FR      |     channel_01    |       60  (=400 x 0.3 / 2)    |   
2020-02-08    |      FR      |     channel_02    |       60  (=400 x 0.3 / 2)    |   
2020-02-08    |      NL      |     channel_01    |       20  (=400 x 0.1 / 2)    |   
2020-02-08    |      NL      |     channel_02    |       20  (=400 x 0.1 / 2)    |  
--------------|--------------|-------------------|-------------------------------|---------
2020-04-15    |      DE      |     channel_01    |       90  (=300 x 0.6 / 2)    |   
2020-04-15    |      DE      |     channel_02    |       90  (=300 x 0.6 / 2)    |   
2020-04-15    |      FR      |     channel_01    |       45  (=300 x 0.3 / 2)    |   
2020-04-15    |      FR      |     channel_02    |       45  (=300 x 0.3 / 2)    |   
2020-04-15    |      NL      |     channel_01    |       15  (=300 x 0.1 / 2)    |   
2020-04-15    |      NL      |     channel_02    |       15  (=300 x 0.1 / 2)    |   

---

在第一个表中，我有costs_per_day_per_country 。
现在，我想根据第二张表中的revenue_share将这些成本分配给每个国家/地区。
此外，分配的成本应平均分配给每个渠道。

我试图用这个查询 go ：

SELECT 
c.event_date, 
c.country, 
c.channel,

  /* Allocate the costs_per_day_per_country according to the revenue share */
  (CASE WHEN c.country IN ('DE') THEN MAX(c.costs_per_day_per_country) * MAX(t1.revenue_share)
        WHEN c.country IN ('FR') THEN MAX(c.costs_per_day_per_country) * MAX(t1.revenue_share)
        WHEN c.country IN ('NL') THEN MAX(c.costs_per_day_per_country) * MAX(t1.revenue_share)
  END) 
  
  /* Split the allocated costs equally to each channel */
  * (1.0 / COUNT(*) OVER (PARTITION BY c.event_date, c.country)) AS costs_allocated
  
FROM costs c
LEFT JOIN

  (SELECT 
  r.event_date, 
  r.country, 
  r.revenue_share
  FROM revenue r
  GROUP BY 1,2,3
  ORDER BY 1,2,3) t1 ON t1.event_date = c.event_date

GROUP BY 1,2,3
ORDER BY 1,2,3;

它几乎给了我预期的结果，但不知何故，它没有正确地将revenue_share应用于每个国家。
它将 0.6 应用于每个国家/地区。

如何修改查询以获得预期结果？

Answer 1

如果我的理解是正确的，那么解决方案比您实现的查询要简单得多，您需要做的就是连接两个表并适当地使用COUNT() over (PARTITION BY)来计算通道数

SELECT 
c.event_date, 
c.country, 
c.channel,
(costs_per_day_per_country * revenue_share)/COUNT(c.channel) over (partition by c.event_date, c.country) dt

FROM costs c
LEFT JOIN
 revenue t1 ON t1.event_date = c.event_date and t1.country=c.country

ORDER BY 1,2,3;