如何在 SQL 中过滤数学计算的数据？

Question

这可能比我想象的要容易，但我很难过。我写了下面的查询，它总结了客户在特定商店的所有购买，以及同一客户在同一商店的退货总额，最后，存储从总退货中减去总退货的结果在名为 TOTAL_SPENT 的计算列中购买。这是我的查询：

SELECT Account_Id,
  SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) AS TOTAL_PURCHASED,
  SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END) AS TOTAL_RETURNED,
  (TOTAL_PURCHASED - TOTAL_RETURNED) AS TOTAL_SPENT
FROM Transactions_Table
WHERE Store = 'Store XYZ'
GROUP BY Account_Id
ORDER BY TOTAL_SPENT DESC

我的问题是这个查询返回了数十万条记录，但我只关心大约 10% 的返回数据。因此，假设您只想查看 TOTAL_SPENT >= 500.00 的返回数据。但是对于我的生活，我无法弄清楚如何做到这一点！ 羞愧地垂下头

Answer 1

在 SQL 中，WHERE 子句在SELECT 子句之前处理。因此，当 SQL 开始过滤表中与查询相关的行时，它还不知道列 TOTAL_SPENT。

有几种方法可以解决这个问题。

1) 在WHERE 子句中重复表达式，而不是在SELECT 子句中引用列名：

SELECT Account_Id,
  SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) AS TOTAL_PURCHASED,
  SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END) AS TOTAL_RETURNED,
  (TOTAL_PURCHASED - TOTAL_RETURNED) AS TOTAL_SPENT
FROM Transactions_Table
WHERE Store = 'Store XYZ'
GROUP BY Account_Id
ORDER BY 
  (SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) - SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END)) DESC

注意：如果您要查找 TOTAL_SPENT 高于 500 的 组，请使用 HAVING 子句而不是 WHERE。

2) 使用视图。删除您的 WHERE 子句并使用您的查询创建一个 VIEW。然后，从该视图中选择并添加您的WHERE 子句。

CREATE VIEW Whatever AS
   SELECT Account_Id,
      SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) AS    TOTAL_PURCHASED,
      SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END) AS TOTAL_RETURNED,
      (TOTAL_PURCHASED - TOTAL_RETURNED) AS TOTAL_SPENT
   FROM Transactions_Table
   GROUP BY Account_Id;

SELECT * From Whatever WHERE TOTAL_SPENT >= 500

3) 使用嵌套的 SELECT：

select * from (
  SELECT
    Account_Id,
    SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) AS TOTAL_PURCHASED,
    SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END) AS TOTAL_RETURNED,
    (TOTAL_PURCHASED - TOTAL_RETURNED) AS TOTAL_SPENT
  FROM Transactions_Table
  WHERE Store = 'Store XYZ'
  GROUP BY Account_Id
) where TOTAL_SPENT > 500 ORDER BY TOTAL_SPENT DESC;

4) 如果您的数据库支持，请使用 CTE：

WITH MyQuery AS (
  SELECT
    Account_Id,
    SUM(CASE WHEN Transaction_Code = 'Purchase' THEN Transaction_Amount ELSE 0 END) AS TOTAL_PURCHASED,
    SUM(CASE WHEN Transaction_Code = 'Return' THEN Transaction_Amount ELSE 0 END) AS TOTAL_RETURNED,
    (TOTAL_PURCHASED - TOTAL_RETURNED) AS TOTAL_SPENT
  FROM Transactions_Table
  WHERE Store = 'Store XYZ'
  GROUP BY Account_Id
)
SELECT * FROM MyQuery WHERE TOTAL_SPENT > 500 ORDER BY TOTAL_SPENT DESC;