我可以在 T-SQL 中使用 while 或 cursor 来插入新的 table

Question

我写了一个查询来获取 personID 的每个日期的总数，我想做的是使用存储过程将结果插入一个新的 table 我的查询类似于参考（使用AdventureWorks2019)

WITH TOP7 AS
(
    SELECT 
        SOH.SalesPersonID, SOH.OrderDate,
        SUM(SOH.TotalDue) AS Total,
        ROW_NUMBER() OVER (PARTITION BY SOH.SalesPersonID ORDER BY SOH.OrderDate DESC) AS RowNum
    FROM 
        Sales.SalesOrderHeader AS SOH
    LEFT JOIN 
        Sales.SalesOrderHeader AS SOH2 ON SOH.SalesOrderID = SOH2.SalesOrderID
    WHERE 
        SOH.SalesPersonID IS NOT NULL 
    GROUP BY
        SOH.OrderDate, SOH.SalesPersonID
)
(
SELECT SalesPersonID, OrderDate, Total, RowNum
FROM TOP7
WHERE RowNum <= 7
)

这是我得到的输出：

ID    Date                        Total      RowNum
----------------------------------------------------
274   2014-05-01 00:00:00.000    42546.9235   1
274   2014-03-31 00:00:00.000   110623.7157   2
274   2014-03-01 00:00:00.000    46525.3068   3
274   2014-01-29 00:00:00.000     1592.5736   4
274   2013-11-30 00:00:00.000    79835.0276   5
274   2013-09-30 00:00:00.000   102227.2339   6
274   2013-08-30 00:00:00.000      2194.914   7
275   2014-05-01 00:00:00.000    417208.4727  1
275   2014-03-31 00:00:00.000    273730.5071  2
275   2014-03-01 00:00:00.000    221438.2289  3
275   2014-01-29 00:00:00.000    279451.2658  4

我想做的是将结果插入一个新的table类似的东西

ID, Fdate, Ftotal, Sdate, Stotal, Tdate, Ttotal, Fdate, Ftotal, Fidate, Fitotal, Sidate, Sitotal, Sedate, Stotal

274 | 2014-05-01 | 42546.9235 | 2014-03-31 | 110623.7157 | 2014-03-01 | 46525.3068  | 2014-01-29 | 1592.5736 | 2013-11-30 | 79835.0276   | 2013-09-30 | 102227.2339 | 2013-08-30 | 2194.914

我正在使用 SQL 服务器。

我想对结果中的所有 ID 执行此操作 - 我该如何实现？

提前致谢

编辑：

我这样做是为了插入新的 table

WITH TOP7 AS
(
    SELECT 
        SOH.SalesPersonID, SOH.OrderDate,
        SUM(SOH.TotalDue) AS Total,
        ROW_NUMBER() OVER (PARTITION BY SOH.SalesPersonID ORDER BY SOH.OrderDate DESC) AS RowNum
    FROM 
        Sales.SalesOrderHeader AS SOH
    LEFT JOIN 
        Sales.SalesOrderHeader AS SOH2 ON SOH.SalesOrderID = SOH2.SalesOrderID
    WHERE 
        SOH.SalesPersonID IS NOT NULL 
    GROUP BY
        SOH.OrderDate, SOH.SalesPersonID
)
    INSERT INTO [Person].[WeekEmployeeTotals] ([PersonID]
      ,[FirstDate]
      ,[FirstDateTotal]
      ,[SecondDate]
      ,[SecondDateTotal]
      ,[ThirdDate]
      ,[ThirdDateTotal]
      ,[FourthDate]
      ,[FourthDateTotal]
      ,[FifthDate]
      ,[FifthDateTotal]
      ,[SixthDate]
      ,[SixthDateTotal]
      ,[SeventhDate]
      ,[SeventhDateTotal])
SELECT  SalesPersonID, 
        Date1 = MAX(CASE WHEN RowNum = 1 THEN OrderDate END),
        Total1 = SUM(CASE WHEN RowNum = 1 THEN Total END),
        Date2 = MAX(CASE WHEN RowNum = 2 THEN OrderDate END),
        Total2 = SUM(CASE WHEN RowNum = 2 THEN Total END),
        Date3 = MAX(CASE WHEN RowNum = 3 THEN OrderDate END),
        Total3 = SUM(CASE WHEN RowNum = 3 THEN Total END),
        Date4 = MAX(CASE WHEN RowNum = 4 THEN OrderDate END),
        Total4 = SUM(CASE WHEN RowNum = 4 THEN Total END),
        Date5 = MAX(CASE WHEN RowNum = 5 THEN OrderDate END),
        Total5 = SUM(CASE WHEN RowNum = 5 THEN Total END),
        Date6 = MAX(CASE WHEN RowNum = 6 THEN OrderDate END),
        Total6 = SUM(CASE WHEN RowNum = 6 THEN Total END),
        Date7 = MAX(CASE WHEN RowNum = 7 THEN OrderDate END),
        Total7 = SUM(CASE WHEN RowNum = 7 THEN Total END)

FROM    TOP7
WHERE   RowNum <= 7
GROUP BY SalesPersonID;

但我有一个问题，如何设置 Select 以插入我的意思是类似的东西

Create PROCEDURE [dbo].[InsertFromSHO]
    @PID INT,
    @D1 datetime,
    @T1 numeric(24,7),
    @D2 datetime,
    @T2 numeric(24,7),
    @D3 datetime,
    @T3 numeric(24,7),
    @D4 datetime,
    @T4 numeric(24,7),
    @D5 datetime,
    @T5 numeric(24,7),
    @D6 datetime,
    @T6 numeric(24,7),
    @D7 datetime,
    @T7 numeric(24,7)

AS
    SET @D1 = (SELECT Date1 = MAX(CASE WHEN RowNum = 1 THEN OrderDate END) FROM TOP7 WHERE RowNum <= 7)

^ 我知道这是错误的，但这是我正在尝试做的，因为我以后需要经常这样做，所以我需要知道一个好的方法来实现它

Answer 1

不需要游标。 SQL 是一种基于集合的语言，在这种语言中，这种操作无需循环即可轻松执行。您可以只使用 insert ... select 语法：

WITH TOP7 AS (
    SELECT 
        SOH.SalesPersonID, SOH.OrderDate,
        SUM(SOH.TotalDue) AS Total,
        ROW_NUMBER() OVER (PARTITION BY SOH.SalesPersonID ORDER BY SOH.OrderDate DESC) AS RowNum
    FROM 
        Sales.SalesOrderHeader AS SOH
    LEFT JOIN 
        Sales.SalesOrderHeader AS SOH2 ON SOH.SalesOrderID = SOH2.SalesOrderID
    WHERE 
        SOH.SalesPersonID IS NOT NULL 
    GROUP BY
)
INSERT INTO mytable(ID, Fdate, Ftotal, Sdate, Stotal, Tdate, Ttotal, Fdate, Ftotal, Fidate, Fitotal, Sidate, Sitotal, Sedate, Stotal)
SELECT SalesPersonID, OrderDate, Total, RowNum
FROM TOP7
WHERE RowNum <= 7

Answer 2

你可以通过条件聚合得到你想要的输出结果

WITH TOP7 AS
(
    SELECT 
        SOH.SalesPersonID, SOH.OrderDate,
        SUM(SOH.TotalDue) AS Total,
        ROW_NUMBER() OVER (PARTITION BY SOH.SalesPersonID ORDER BY SOH.OrderDate DESC) AS RowNum
    FROM 
        Sales.SalesOrderHeader AS SOH
    LEFT JOIN 
        Sales.SalesOrderHeader AS SOH2 ON SOH.SalesOrderID = SOH2.SalesOrderID
    WHERE 
        SOH.SalesPersonID IS NOT NULL 
    GROUP BY
        SOH.OrderDate, SOH.SalesPersonID
)
SELECT  SalesPersonID, 
        Date1 = MAX(CASE WHEN RowNum = 1 THEN OrderDate END),
        Total1 = SUM(CASE WHEN RowNum = 1 THEN Total END),
        Date2 = MAX(CASE WHEN RowNum = 2 THEN OrderDate END),
        Total2 = SUM(CASE WHEN RowNum = 2 THEN Total END),
        Date3 = MAX(CASE WHEN RowNum = 3 THEN OrderDate END),
        Total3 = SUM(CASE WHEN RowNum = 3 THEN Total END),
        Date4 = MAX(CASE WHEN RowNum = 4 THEN OrderDate END),
        Total4 = SUM(CASE WHEN RowNum = 4 THEN Total END),
        Date5 = MAX(CASE WHEN RowNum = 5 THEN OrderDate END),
        Total5 = SUM(CASE WHEN RowNum = 5 THEN Total END),
        Date6 = MAX(CASE WHEN RowNum = 6 THEN OrderDate END),
        Total6 = SUM(CASE WHEN RowNum = 6 THEN Total END),
        Date7 = MAX(CASE WHEN RowNum = 7 THEN OrderDate END),
        Total7 = SUM(CASE WHEN RowNum = 7 THEN Total END)
FROM    TOP7
WHERE   RowNum <= 7
GROUP BY SalesPersonID;

Example on DB<>Fiddle

不过，以这种格式存储数据可能不是一个好主意，如果您需要经常访问这种格式的数据，最好创建一个视图。我也删除了你在这里的加入，因为它似乎完全没有意义，因为你从不引用 SOH2:

CREATE VIEW Top7SalesPivot
AS
WITH TOP7 AS
(
    SELECT 
        SOH.SalesPersonID, 
        SOH.OrderDate,
        SUM(SOH.TotalDue) AS Total,
        ROW_NUMBER() OVER (PARTITION BY SOH.SalesPersonID ORDER BY SOH.OrderDate DESC) AS RowNum
    FROM 
        Sales.SalesOrderHeader AS SOH
    WHERE 
        SOH.SalesPersonID IS NOT NULL 
    GROUP BY
        SOH.OrderDate, SOH.SalesPersonID
)
SELECT  SalesPersonID, 
        Date1 = MAX(CASE WHEN RowNum = 1 THEN OrderDate END),
        Total1 = SUM(CASE WHEN RowNum = 1 THEN Total END),
        Date2 = MAX(CASE WHEN RowNum = 2 THEN OrderDate END),
        Total2 = SUM(CASE WHEN RowNum = 2 THEN Total END),
        Date3 = MAX(CASE WHEN RowNum = 3 THEN OrderDate END),
        Total3 = SUM(CASE WHEN RowNum = 3 THEN Total END),
        Date4 = MAX(CASE WHEN RowNum = 4 THEN OrderDate END),
        Total4 = SUM(CASE WHEN RowNum = 4 THEN Total END),
        Date5 = MAX(CASE WHEN RowNum = 5 THEN OrderDate END),
        Total5 = SUM(CASE WHEN RowNum = 5 THEN Total END),
        Date6 = MAX(CASE WHEN RowNum = 6 THEN OrderDate END),
        Total6 = SUM(CASE WHEN RowNum = 6 THEN Total END),
        Date7 = MAX(CASE WHEN RowNum = 7 THEN OrderDate END),
        Total7 = SUM(CASE WHEN RowNum = 7 THEN Total END)
FROM    TOP7
WHERE   RowNum <= 7
GROUP BY SalesPersonID;

这样您就不需要在每次添加销售时都更新 table。

如果效果不佳，您可能会发现以下索引有帮助：

CREATE NONCLUSTERED INDEX IX_Sales_SalesOrderHeader__SalesPersonID_OrderDate 
    ON Sales.SalesOrderHeader (SalesPersonID, OrderDate) 
    INCLUDE (TotalDue);

我可以在 T-SQL 中使用 while 或 cursor 来插入新的 table

Can I use while or cursor in T-SQL to insert into new table

tsql

sql-server

loops

stored-procedures

sql-insert