SQL 查询两个 PIVOT 列组合

Posted 2023-03-31

【中文标题】SQL 查询两个 PIVOT 列组合

【英文标题】：SQL query two PIVOT columns combine

【发布时间】：2020-06-19 09:50:08

【问题描述】：

select pvtMonth.CardName, [1] as [Jan Sales], [2] as [Feb Sales], [3] as [Mar Sales], [4] as [Apr Sales], [5] as [May Sales], 
[6] as [Jun Sales], [7] as [Jul Sales], [8] as [Aug Sales], [9] as [Sep Sales], [10] as [Oct Sales], [11] as [Nov Sales], [12] as [Dec Sales] from 
(
select X.CardName, SUM(X.[Total Sales S$]) as [Sales $] , X.Month from Data X group by X.CardName ,X.Month
) X PIVOT 
(
    sum(X.[Sales $])
    FOR [Month]
    IN ( [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12] )
) AS pvtMonth order by pvtMonth.CardName asc 

现在我需要再添加一个名为“Jan Gp”的列名是“X.Gp”。如何更新我的查询。我需要如下结果集

【问题讨论】：

SQL Server，还是别的什么？另外，我觉得这有很多重复；你先搜索了吗？

【参考方案1】：

我建议使用条件聚合而不是 pivot 语法。它更加灵活（并且至少同样有效）。你似乎想要这样的东西：

select 
    cardName,
    sum(case when month = 1 then [Total Sales S$]  end) [Jan Sales $],
    sum(case when month = 1 then [Total Sales SGP] end) [Jan Sales GP],
    sum(case when month = 2 then [Total Sales S$]  end) [Feb Sales $],
    sum(case when month = 2 then [Total Sales SGP] end) [Feb Sales GP],
    ...
from data
group by cardName

【讨论】：

您知道，“条件聚合”被称为“CROSSTAB”。

@JeffModen 怎么做可以帮忙？

@GMB 上面的回答是正确的。它使用 CROSSTAB，它通常比 PIVOT 更快（尤其是对于您要求的两列 Pivots），允许进行更多的自定义，并且可以轻松实现行总计和列总计。