Question

“交易”表是包含date，customerID，每单位费用（price）和quantity的所有交易的列表。

我需要一份报告，将交易分组为几个月，并计算每个客户的总收入和平均收入。

我无法弄清楚如何将group by customerID插入此select语句中：

SELECT extract(month FROM date) month,
       count(*) purchases,
       SUM(price*quantity) income,
       AVG(SUM(price*quantity)) <-- this needs to be grouped by customerID aswell as month
FROM
    transactions
WHERE
    date BETWEEN i2 AND i3
GROUP BY extract(month FROM date);

我能得到的最接近的是：

SELECT extract(month FROM date) month,
       count(*) purchases,
       SUM(price*quantity) income,
       ( SELECT AVG(SUM(price*quantity))
         FROM transactions
         GROUP BY customerID, extract(month FROM date) )
FROM
    transactions t
WHERE
    date BETWEEN i2 AND i3
GROUP BY extract(month FROM date);

但是，由于嵌套的select将返回多行，因此在我考虑它一分钟后，这种方法就停止了。如果我添加where子句并删除第二个选择中的group by date：

(SELECT AVG(SUM(price*quantity))
 FROM transactions
 GROUP BY customerID
 where extract(month FROM date) = month)

然后它会回来......我不确定是什么，但这是错误的。

顺便说一句，它是一个oracle数据库（我认为是10g）。

Answer 1

您可以使用算术来完成此操作。。。将总数除以客户数量：

SELECT extract(month FROM date) month,
       count(*) purchases,
       SUM(price*quantity) income,
       sum(price*quantity) / count(distinct customerId)
FROM
    transactions
WHERE
    date BETWEEN i2 AND i3
GROUP BY extract(month FROM date);

客户每月的SQL平均值

1 个答案: