需要一个 postgresql 查询来分组和排序

Question

我有一个如下的客户表：

customername, ordername, amount
=============================
bob, book,     20
bob, computer, 40
steve,hat,     15
bill, book,    12
bill, computer, 3
steve, pencil, 10
bill, pen,      2

我想运行查询以获得以下结果：

customername, ordername, amount
=============================
bob,   computer, 40
bob,   book,     20
bob,   ~total~,  60
steve, hat,      15
steve, pencil, 10
steve, ~total~,25 
bill, book,    12
bill, computer, 3
bill, pen,      2 
bill, ~total~, 17

我希望从最大到最小订购每个客户的金额，并将新的订单名称命名为“~total~”（必须始终是每个客户的最后一行），结果是同一客户的所有金额的总和. 所以，在上面的例子中，bob 应该是第一个，因为总 = 60，史蒂夫第二（总 = 25）和第三个（总 = 17）。

Answer 1

用途：

SELECT x.customername,
       x.ordername,
       x.amount
  FROM (SELECT a.customername,
               a.ordername,
               a.amount,
               y.rk,
               1 AS sort
          FROM CUSTOMERS a
          JOIN (SELECT c.customername,
                       ROW_NUMBER() OVER (ORDER BY SUM(c.amount) DESC) AS rk
                  FROM CUSTOMERS c
              GROUP BY c.customername) y ON y.customername = a.customername
        UNION ALL
        SELECT b.customername,
               '~total~',
               SUM(b.amount),
               ROW_NUMBER() OVER (ORDER BY SUM(b.amount) DESC) AS rk,              
               2 AS sort
          FROM CUSTOMERS b
      GROUP BY b.customername) x
ORDER BY x.rk, x.customername, x.sort, x.amount DESC

您可以查看使用 GROUP BY ROLLUP，但 ordername 的值为 NULL，因此您必须对其进行后处理才能将其替换为“~total~”...