Question

我有一个表，其中包含每日库存数据，我想按周进行汇总。每日餐桌是：

CREATE TABLE [dbo].[TempDaily] (
    [Symbol]    CHAR (10)  NOT NULL,
    [CloseDate] DATE       NOT NULL,
    [DailyHi]   FLOAT (53) NULL,
    [DailyLow]  FLOAT (53) NULL,
    [AdjClose]  FLOAT (53) NOT NULL,
    [WeekEnd]   DATE       NULL
);

我想在每周表中插入一周的摘要：

CREATE TABLE [dbo].[Weekly] (
    [Symbol]    CHAR (10)  NOT NULL,
    [WeekEnd]   DATE       NOT NULL,
    [WeeklyHi]  FLOAT (53) NOT NULL,
    [WeeklyLow] FLOAT (53) NOT NULL,
    [AdjClose]  FLOAT (53) NULL
);

[AdjClose]列设置为允许为空，因为我当前的解决方法是先插入其他4列，然后使用来自第三张表的每周AdjClose值更新每周表，这确实很慢。

将每周的前四列数据放入表中很简单：

strSQL = "INSERT INTO Weekly (Symbol, WeekEnd, WeeklyHi, WeeklyLow) " &
    "SELECT Symbol, WeekEnd, MAX(DailyHi), MIN(DailyLow) " &
     "FROM TempDaily " &
     "GROUP BY Symbol, WeekEnd "

获取每周AdjClose并不是那么简单，但是我可以将数据插入临时表中，然后从中更新“每周”表：

strSQL = "INSERT INTO Test (WeekEnd, AdjClose) " &
    "Select wdata.WeekEnd, MAX(wdata.AdjClose) " &
    "FROM " &
        "(Select CloseDate, WeekEnd, " &
        "FIRST_VALUE(AdjClose) OVER (PARTITION BY WeekEnd ORDER BY CloseDate                                            
         DESC ROWS UNBOUNDED PRECEDING) As AdjClose " &
         "FROM TempDaily) wdata " &
     "GROUP BY wdata.WeekEnd "

我真的更喜欢用一条语句将所有5列的数据插入到Weekly表中，而不是依靠笨拙的解决方法，但是我一直无法弄清楚如何将2条语句联接起来。

“每日”表中的数据采用以下格式（为清楚起见添加了空格）：

Symbol,    CloseDate,  DailyHi,DailyLow,AdjClose,   WeekEnd
AAPL      ,5/31/2019,   177.99,  175.00,  175.07,  6/1/2019
AAPL      ,5/30/2019,   179.23,  176.67,  178.30,  6/1/2019
AAPL      ,5/29/2019,   179.35,  176.00,  177.38,  6/1/2019
AAPL      ,5/28/2019,   180.59,  177.91,  178.23,  6/1/2019
AAPL      ,5/24/2019,   182.14,  178.62,  178.97,  5/25/2019
AAPL      ,5/23/2019,   180.54,  177.81,  179.66,  5/25/2019
AAPL      ,5/22/2019,   185.71,  182.55,  182.78,  5/25/2019
AAPL      ,5/21/2019,   188.00,  184.70,  186.60,  5/25/2019
AAPL      ,5/20/2019,   184.35,  180.28,  183.09,  5/25/2019

每周表格的结尾应为：

Symbol,        WeekEnd, WeeklyHi, WeeklyLo,  AdjClose
AAPL      ,   6/1/2019,   180.59,   175.00,   175.07
AAPL      ,   5/25/2019,  188.00,   177.81,   178.97

我正在将MS Access数据库转换为VB.NET，并认为这将是一个相当简单的端口。差异比我预期的要多得多。任何帮助表示赞赏。

下面介绍的解决方案是我尝试过的一种解决方案，但它会生成一个SQLException：''（'附近的语法不正确。'wdata'附近的语法不正确。所以我仍在尝试解决这一问题。

strSQL = "INSERT INTO Weekly (Symbol, WeekEnd, WeeklyHi, WeeklyLow, AdjClose) " &
"Select A.Symbol, A.WeekEnd, A.WeeklyHi, A.WeeklyLow, ISNULL(B.AdjClose, 0) as AdjClose " &
"FROM " &
    "(SELECT Symbol, WeekEnd, MAX(DailyHi) as WeeklyHi, MIN(DailyLow) as WeeklyLow " &
     "FROM TempDaily " &
     "GROUP BY Symbol, WeekEnd ) A " &
     "LEFT JOIN " & 
    "(Select wdata.WeekEnd, MAX(wdata.AdjClose) as AdjClose" &
    "FROM " &
        "(Select CloseDate, WeekEnd, " &
        "FIRST_VALUE(AdjClose) OVER (PARTITION BY WeekEnd ORDER BY CloseDate                                            
         DESC ROWS UNBOUNDED PRECEDING) As AdjClose " &
         "FROM TempDaily) wdata " &
     "GROUP BY wdata.WeekEnd) B ON A.WeekEnd = B.WeekEnd "

Answer 1

我认为您可以做自己想做的事情，像这样（更新）加入他们：

    strSQL = "INSERT INTO Weekly (Symbol, WeekEnd, WeeklyHi, WeeklyLow, AdjClose) " &
"Select A.Symbol, A.WeekEnd, A.WeeklyHi, A.WeeklyLow, ISNULL(B.AdjClose, 0) as AdjClose " &
"FROM " &
    "(SELECT Symbol, WeekEnd, MAX(DailyHi) as WeeklyHi, MIN(DailyLow) as WeeklyLow " &
    "FROM TempDaily " &
    "GROUP BY Symbol, WeekEnd ) A " &
    "LEFT JOIN " &
    "(Select wdata.WeekEnd, MAX(wdata.AdjClose) as AdjClose " &
    "FROM " &
        "(Select CloseDate, WeekEnd, " &
        "FIRST_VALUE(AdjClose) OVER (PARTITION BY WeekEnd ORDER BY CloseDate " &
        "DESC ROWS UNBOUNDED PRECEDING) As AdjClose " &
        "FROM TempDaily) wdata " &
     "GROUP BY wdata.WeekEnd) B ON A.WeekEnd = B.WeekEnd "

如何在2个带有GROUP的SELECT语句之间创建JOIN？

1 个答案: