在单个查询中聚合时间范围和聚合所有时间

Question

我被要求编写一个与 MySQL/PostgreSQL 兼容的单一查询，该查询显示销售代表（sales_rep 此处）在给定月份的总销售额，并在另一列中显示同一个人全年的平均月销售额。到目前为止我提出的查询如下：

SELECT sr.name, SUM(s.amount) as total, ??
  FROM sales_rep as sr, sales as s
    WHERE sr.id = s.sales_rep_id
      AND s.date >= '2000-12-01' 
      AND s.date < '2001-01-01'
  GROUP BY sr.name
  ORDER BY total DESC 
LIMIT 10;

你可以在这里试试：http://sqlfiddle.com/#!17/67c65/10

Answer 1

你可以使用窗口函数：

SELECT YEAR(s.date), MONTH(s.date), sr.name, SUM(s.amount) as total, 
       SUM(SUM(s.amount)) OVER () / 12 as monthly_average
FROM sales_rep sr JOIN
     sales s
     ON sr.id = s.sales_rep_id
WHERE s.date >= '2000-12-01' AND s.date < '2001-01-01'
GROUP BY sr.name, YEAR(s.date), MONTH(s.date);

注意：这会将没有销售的月份计算为 0。如果您想忽略它们：

SELECT YEAR(s.date), MONTH(s.date), sr.name, SUM(s.amount) as total, 
       AVG(SUM(s.amount)) OVER () as monthly_average
FROM sales_rep sr JOIN
     sales s
     ON sr.id = s.sales_rep_id
WHERE s.date >= '2000-12-01' AND s.date < '2001-01-01'
GROUP BY sr.name, YEAR(s.date), MONTH(s.date);

Answer 2

SELECT sr.name, SUM(s.amount) as total, avg(s.amount) as average
FROM sales_rep as sr
join sales as s
on sr.id = s.sales_rep_id
where s.date >= '2000-12-01' 
AND s.date < '2001-01-01'
GROUP BY sr.name
LIMIT 100;

请使用显式连接。 这意味着你将改变这一点：

FROM sales_rep as sr, sales as s
WHERE sr.id = s.sales_rep_id

用这个：

FROM sales_rep as sr
join sales as s
on sr.id = s.sales_rep_id

最后，在您限制前 100 个之前，按数量或平均值而不是按名称排序是有意义的。因此，这个解决方案：

SELECT sr.name as name, SUM(s.amount) as total, avg(s.amount) as average
FROM sales_rep as sr, sales as s
WHERE sr.id = s.sales_rep_id
AND s.date >= '2000-12-01' 
AND s.date < '2001-01-01'
GROUP BY sr.name
order by total
limit 100;