问题描述
真的很纠结这个。寻找一种查询不同时间跨度的销售量的方法，所有这些都放在一个表中。
假设现在是 8 月 13 日。然后我想总结以下销售量：
- 7 月 1 日 00:00:00 - 7 月 31 日 11:59:59（过去 1 个月）
- 5 月 - 1 日 00:00:00 - 7 月 31 日 11:59:59（过去 3 个月）
- 2 月 1 日 00:00:00 - 7 月 31 日 11:59:59（过去 6 个月）
- 8 月 1 日 00:00:00 - 7 月 31 日 11:59:59（过去 12 个月）
我尝试重用数据透视查询示例中的片段，这些片段可以找到与销售量相关但没有成功。 我从头开始，首先尝试创建可用于时间跨度的日期时间，但这已经非常繁琐，而且不知何故感觉不对。
我认为我需要以 4 个单独的查询结束，然后需要将这些查询连接起来以将所有列合并到一个表中：
| 物品名称 | 项目代码 | Sal.vol 1 月 | Sal.vol 3 个月 | Sal.vol 6 个月 | Sal.vol 12 个月 |
|---|---|---|---|---|---|
| 项目 1 | 10102251 | 10 | 30 | 60 | 120 |
| 项目 2 | 10120101 | 14 | 35 | 78 | 181 |
我目前的进展：
SELECT
T1.itemcode,month(T1.[docDate]) as month,T3.ItemName,SUM(t1.quantity )
FROM INV1 T1 INNER JOIN OINV T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between DATEADD(month,-4,GETDATE()) and DATEADD(month,-1,GETDATE()) AND T1.ItemCode LIKE '1%' AND T3.ItmsGrpCod = 100
Group By T1.itemcode,month(T1.[docDate])
对于我的约会时间，我想按如下方式构建它们：
DECLARE @date date = month(DATEADD(month,GETDATE())) + '-01-' + year(DATEADD(month,GETDATE()));
DECLARE @datetime datetime = @date;
SELECT @datetime AS '@datetime'
但是我还没有成功，因为我在 int 到日期转换方面遇到了麻烦，我认为 Jan-Sep 的月份（）将不起作用，因为它只返回 1 位数字...
有人能帮我指出正确的方向吗？
提前致谢！
更新 1：
@imerd 为我指明了正确的方向。我能够根据需要构建我的查询。分享我的中间结果，因为这现在有效。 Query 需要一些清理和优化，重复出现的主要部分外包到第二个函数中。
create function MonthFirst (@period integer ) returns date
as begin
declare @returnperiod date
set @returnperiod = cast(cast(year(DATEADD(month,@period,GETDATE())) as varchar(4)) + '-' +
cast(month(DATEADD(month,GETDATE())) as varchar(2)) + '-01' as date)
return @returnperiod
end
-- drop function dbo.MonthFirst
SELECT S12.ItemCode,o1.ItemName,S12.[last 12 months],S6.[last 6 months],S3.[last 3 months],S1.[last month] FROM
(SELECT
T1.itemcode,CAST(SUM(t1.quantity) AS int) AS 'last 12 months'
FROM DLN1 T1 INNER JOIN ODLN T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between CAST(dbo.MonthFirst(-12) AS datetime) AND DATEADD(second,CAST(dbo.MonthFirst(0) AS datetime))
AND T3.validFor = 'Y' AND T3.ItmsGrpCod = 100 --products / 102 raw materials
AND T2.CANCELED = 'N'
Group By T1.itemcode) S12
FULL JOIN
(SELECT
T1.itemcode,CAST(SUM(t1.quantity) AS int) AS 'last 6 months'
FROM DLN1 T1 INNER JOIN ODLN T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between CAST(dbo.MonthFirst(-6) AS datetime) AND DATEADD(second,CAST(dbo.MonthFirst(0) AS datetime))
AND T3.validFor = 'Y' AND T3.ItmsGrpCod = 100 --products / 102 raw materials
AND T2.CANCELED = 'N'
Group By T1.itemcode) S6 on S12.ItemCode = S6.ItemCode
FULL JOIN
(SELECT
T1.itemcode,CAST(SUM(t1.quantity) AS int) AS 'last 3 months'
FROM DLN1 T1 INNER JOIN ODLN T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between CAST(dbo.MonthFirst(-3) AS datetime) AND DATEADD(second,CAST(dbo.MonthFirst(0) AS datetime))
AND T3.validFor = 'Y' AND T3.ItmsGrpCod = 100 --products / 102 raw materials
AND T2.CANCELED = 'N'
Group By T1.itemcode) S3 on S12.ItemCode = S3.ItemCode
FULL JOIN
(SELECT
T1.itemcode,CAST(SUM(t1.quantity) AS int) AS 'last month'
FROM DLN1 T1 INNER JOIN ODLN T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between CAST(dbo.MonthFirst(-1) AS datetime) AND DATEADD(second,CAST(dbo.MonthFirst(0) AS datetime))
AND T3.validFor = 'Y' AND T3.ItmsGrpCod = 100 --products / 102 raw materials
AND T2.CANCELED = 'N'
Group By T1.itemcode) S1 on S12.ItemCode = S1.ItemCode
INNER JOIN OITM o1 on S12.ItemCode = o1.ItemCode
ORDER BY o1.ItemName
更新 2：
我现在结合了来自@imerd 和@Gordon Linoff 的建议/提示，得到了这个漂亮的小查询，这几乎正是我想要的。
非常感谢！
这是我的解决方案：
DECLARE @period integer = -12
DECLARE @startperiod date
DECLARE @endperiod date
SET @startperiod = CAST(cast(year(DATEADD(month,GETDATE())) AS varchar(4)) + '-' +
cast(month(DATEADD(month,GETDATE())) AS varchar(2)) + '-01' AS date)
SET @endperiod = CAST(cast(year(DATEADD(month,GETDATE())) AS varchar(2)) + '-01' AS datetime)
SELECT T1.ItemCode,sum(case when datediff(month,T1.[docDate],GETDATE()) <= 12 then t1.quantity else 0 end) as month_12,GETDATE()) <= 6 then t1.quantity else 0 end) as month_6,GETDATE()) <= 3 then t1.quantity else 0 end) as month_3,GETDATE()) = 1 then t1.quantity else 0 end) as month_1
FROM DLN1 T1 INNER JOIN ODLN T2 ON T1.docentry = T2.docentry INNER JOIN OITM T3 ON T1.ItemCode = T3.ItemCode
WHERE T1.[docDate] between CAST(@startperiod AS datetime) AND DATEADD(second,CAST(@endperiod AS datetime))
AND T3.validFor = 'Y' AND T3.ItmsGrpCod = 100 --products / 102 raw materials
AND T2.CANCELED = 'N'
GROUP BY T1.ItemCode,T3.ItemName
ORDER BY T3.ItemName ASC
@Gordon Linoff: '=' 需要是 '
再次感谢您的快速帮助！
解决方法
这看起来像一个简单的条件聚合示例：
select itemname,itemcode
sum(case when datediff(month,date,1) = 1 then quantity else 0 end) as month_1,sum(case when datediff(month,1) <= 3 then quantity else 0 end) as month_3,1) <= 6 then quantity else 0 end) as month_6,1) <= 12 then quantity else 0 end) as month_12
from t
group by itemname,itemcode;
您的查询比您的描述所暗示的要复杂，但这应该会告诉您如何进行问题核心的计算。
请注意，datediff() 计算两个日期之间边界的数量，因此非常适用于日历月。
转换需要调整将所有输入转换为单一数据类型，然后转换为日期数据类型 -
对于美式日期格式
declare @date date = cast( cast(month(DATEADD(month,@period,GETDATE())) as varchar(2)) + '-01-'
+ cast(year(DATEADD(month,GETDATE())) as varchar(4)) as date)
对于 ISO 日期格式
declare @date date = cast(cast(year(DATEADD(month,-1,GETDATE())) as varchar(4)) + '-' +
cast(month(DATEADD(month,GETDATE())) as varchar(2))
+ '-01' as date)
假设 ISO 格式是我的小脑子，我从来不明白为什么 Mon DD YYYY 对演示以外的数据处理很实用
创建一个可以传递“要报告的月份数”的函数 - 以减少主报告查询中的代码
create function MonthFirst (@period integer ) returns date
as begin
declare @returnperiod date
set @returnperiod = cast(cast(year(DATEADD(month,GETDATE())) as varchar(2))
+ '-01' as date)
return @returnperiod
end
然后根据需要在查询中使用函数（例如）
select [3 Months Ago] = dbo.MonthFirst(-2) -- 2020-12-01
select [Last Month] = dbo.MonthFirst(-1) -- 2021-02-01
select [This Month] = dbo.MonthFirst(0) -- 2021-03-01
select [Next Month] = dbo.MonthFirst(1) -- 2021-04-01