2

我有一张这样的桌子：

Date        Sales   Department Store
02/01/2012   4.09    Toys       A
03/01/2012   5.90    Toys       A
02/01/2012   6.64    Toys       B
03/01/2012   4.71    Toys       B
02/01/2012   16.72   Toys       C
03/01/2012   1.37    Toys       C
02/01/2012   13.22   Movies     A
03/01/2012   18.06   Movies     A
02/01/2012   5.97    Movies     B
03/01/2012   16.71   Movies     B
02/01/2012   15.38   Movies     C
03/01/2012   19.75   Movies     C

并想要一个这样的表，其中只考虑存储 A 和 B：

Date        Toys    Movies
02/01/2012   10.73   19.19 
03/01/2012   10.61   34.77

这里是我们将在 EXCEL 中使用的 SUMIFS 函数：

=SUMIFS(Value;Date;$H4;Department;I$3;Store;"<>C")

我们可以用 SQL 写什么？

考虑到这是一个简短的例子，数据库表有 30 多个部门和更多的日期。我正在使用脚本

SELECT DISTINCT Date,
ISNULL(MAX(CASE WHEN Department = 'Toys'  AND Store = 'A' THEN Sales END),0) +
ISNULL(MAX(CASE WHEN Department = 'Toys'  AND Store = 'B' THEN Sales END),0) [Toys],
ISNULL(MAX(CASE WHEN Department = 'Movies'  AND Store = 'A' THEN Sales END),0) +
ISNULL(MAX(CASE WHEN Department = 'Movies'  AND Store = 'B' THEN Sales END),0) [Movies]
FROM Table$
GROUP BY Date
ORDER BY Date

...但效率不高。感谢您的任何提示。

4

1 回答 1

2

您的查询没问题，但可以稍微改进一下：

SELEC Date,
      MAX(CASE WHEN Department = 'Toys' THEN Sales else 0 END) as [Toys],
      MAX(CASE WHEN Department = 'Movies' THEN Sales else 0 END) as [Movies]
FROM Table$
WHERE store in ('A', 'B')
GROUP BY Date
ORDER BY Date;

这将删除distinct，这对于 a 是不必要的group by。它将条件移至store子句where，因为它适用于所有行。并且，它通过在语句中ISNULL()包含来删除——因此该部门中没有销售的商店将返回 a而不是.else 0case0NULL

于 2013-06-25T10:45:09.477 回答