Question

我有一张包含以下数据的表格：

Month|Order|OrdersAcc
APR  |4    |4
MAY  |3    |7

我需要的是从五月而不是四月开始累积，这意味着结果应该是：

Month|Order|OrdersAcc
APR  |4    | NULL
MAY  |3    |4
JUN  |2    |7
JUL  |4    |9

请协助查询以实现此目的。

让我简化一下。这是执行累积的查询：

  SELECT        SUM(t2.OrdersAcc) AS OrdersAcc, t1.MONTH_sort2, t1.ActMonth

FROM            dbo.tbl_ME_OrderCompletions AS t1 INNER JOIN

                             (SELECT        SUM(Orders) AS OrdersAcc, Region, MONTH_sort2

                               FROM            dbo.tbl_ME_OrderCompletions

                               GROUP BY Region, MONTH_sort2, ActMonth) AS t2 ON t1.Region = t2.Region AND t1.MONTH_sort2 >= t2.MONTH_sort2

GROUP BY t1.Region, t1.MONTH_sort2, ActMonth

我正在使用month_sort2列，因为它有从1到12的增量整数，1表示4月，2表示5月，依此类推。第一个月（四月）的数据需要在5月份显示，然后从五月开始需要累积。我希望这是有帮助的。

Answer 1

如果您使用的是SQL Server 2012或更高版本，我建议您查看窗口函数 - 尤其是LEAD＆amp; LAG。

还有其他方法可以做到这一点，但它展示了窗口函数的强大功能以及如何将它们链接在一起。

CREATE TABLE dbo.tbl_ME_OrderCompletions
(   
    Month_Sort2 INT
   ,Orders INT
   ,Region NVARCHAR(20)
)

INSERT INTO dbo.tbl_ME_OrderCompletions
VALUES (1,4,'Placeholder')
      ,(2,3,'Placeholder')
      ,(3,2,'Placeholder')
      ,(4,4,'Placeholder')
      ,(5,3,'Placeholder')
      ,(6,1,'Placeholder')
      ,(7,2,'Placeholder')
      ,(8,4,'Placeholder')
      ,(9,3,'Placeholder')
      ,(10,7,'Placeholder')
      ,(11,3,'Placeholder')
      ,(12,9,'Placeholder')


SELECT UPPER(CONVERT(NVARCHAR(3),DATENAME(mm,DATEADD(mm,t.MonthID + 3,-1)))) AS Month
      ,t.Orders
      ,LAG(t.SumOrders) OVER (ORDER BY MonthID) AS OrdersAcc
FROM
     (
        SELECT o.Month_Sort2 AS MonthID
              ,o.Orders AS Orders
              ,SUM(o.Orders) OVER (ORDER BY o.Month_Sort2) AS SumOrders
        FROM dbo.tbl_ME_OrderCompletions o
     ) t

Answer 2

您可以尝试以下操作：

DECLARE @DataSource TABLE
(
    [month_row_id] INT
   ,[motnh] VARCHAR(12)
   ,[order] INT

);

INSERT INTO @DataSource ([month_row_id], [motnh], [order])
VALUES (1, 'APR', 4)
      ,(2, 'MAY', 3)
      ,(3, 'JUN', 2)
      ,(4, 'JULY',4);

SELECT [motnh]
      ,[order]
      ,SUM([order]) OVER (ORDER BY [month_row_id] ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) - [order] AS [OrderAcc]
FROM @DataSource;

我们的想法是将值累积到当前行，但使用简单的-操作排除当前行值。

另请注意，我添加了month_row_id列，以便能够及时订购行，因为使用月份名称无效。

Answer 3

我找到了解决方案。

SELECT        SUM(t2.OrdersAcc) AS OrdersAcc, t1.MONTH_sort2, t1.ActMonth

FROM            dbo.tbl_ME_OrderCompletions AS t1 INNER JOIN

                             (SELECT        SUM(Orders) AS OrdersAcc, Region, MONTH_sort2

                               FROM            dbo.tbl_ME_OrderCompletions


                               GROUP BY Region, MONTH_sort2, ActMonth) AS t2 ON t1.MONTH_sort2 > t2.MONTH_sort2

GROUP t1.MONTH_sort2, ActMonth

我所要做的就是将Month_sort2连接从等于或大于大于。谢谢你的帮助

Answer 4

你需要澄清一下你的问题。但是从我收集的内容来看，你正在寻找：

; WITH t AS (
  SELECT 4 AS monthNum, 'APR' AS m, 4 AS t UNION ALL
  SELECT 5 AS monthNum, 'MAY' AS m, 3 AS t UNION ALL
  SELECT 6 AS monthNum, 'JUN' AS m, 2 AS t UNION ALL
  SELECT 7 AS monthNum, 'JUL' AS m, 4 AS t
)
SELECT s1.m,s1.t
    , SUM(s1.prevT) OVER (ORDER BY s1.monthNum ROWS UNBOUNDED PRECEDING) AS totalT
FROM (
    SELECT t.monthNum, t.m, t.t
        , (LAG(t.t) OVER (ORDER BY t.monthNum)) AS prevT 
    FROM t
) s1
;

1）这仅适用于MS SQL Server 2012或更高版本。

2）你必须有办法订购你的月份。我添加了一个monthNum列来执行此操作。

=============================================== ====

对于SQL 2008和2008R2：

; WITH t AS (
    SELECT 4 AS monthNum, 'APR' AS m, 4 AS t UNION ALL
    SELECT 5 AS monthNum, 'MAY' AS m, 3 AS t UNION ALL
    SELECT 6 AS monthNum, 'JUN' AS m, 2 AS t UNION ALL
    SELECT 7 AS monthNum, 'JUL' AS m, 4 AS t
)
SELECT t3.m, t3.t, NULLIF(t3.runningTotal, t3.t) AS nulledRunningTotal
FROM (
    SELECT t.monthNum, t.m, t.t
        , ( SELECT SUM(t2.t) FROM t t2 WHERE t2.monthNum <= t.monthNum ) AS runningTotal
    FROM t
) t3 ;

我想这可能是我第一次使用NULLIF()。因此，如果您需要在2008年之前，可以将NULLIF(t3.runningTotal,t3.t)更改为CASE WHEN t3.runningTotal<>t3.t THEN t3.runningTotal ELSE NULL END。

从下个月开始累积

4 个答案: