如何在 SQL 上使用 PIVOT 以及列数检查和分组等附加条件？

Question

下面 SQL 查询的当前输出是：

Time | SumValue(CE+PE)

2 列 - 确保输出的数量正好是十。

如何将其转置为：

Time | CE1Value | CE2Value |... |CE5Value | PE1Value | ... PE5Value

11 列。

确保同一特定时间的十条记录全部存在，否则跳到下一分钟（下一条记录）。我假设它是按时间排序的，默认情况下我可以使用 order by - 它是最佳的吗？

针对 MS Access 中的数据使用 windows 中的独立 vbs 文件。

mySQL = "SELECT [Data.Time],"& Quantity &"*Sum([Data.Price]) AS SumOfPrice FROM Data WHERE (  ( [Data.Ticker] Like '"& CE1 &"'  Or [Data.Ticker] Like '"& CE2 &"' Or [Data.Ticker] Like '"& CE3 &"' Or [Data.Ticker] Like '"& CE4 &"' Or [Data.Ticker] Like '"& CE5 &"' Or [Data.Ticker] Like '"& PE1 &"'  Or [Data.Ticker] Like '"& PE2 &"'  Or [Data.Ticker] Like '"& PE3 &"'  Or [Data.Ticker] Like '"& PE4 &"'  Or [Data.Ticker] Like '"& PE5 &"'  ) AND ([Data.DateTr]=#"& DateIn &"#)) GROUP BY [Data.Time] HAVING [Data.Time] > #"& startTime(i) &"# and (((Count([Data.Ticker]))=10))"

例如。为了理解：数据table是：

DateTr|Ticker|Time|Price
22July2016|Apple|3:20 AM|15
22July2016|Apple|3:21 AM|16
22July2016|Apple|3:23 AM|14
22July2016|Amazon|3:20 AM|16
22July2016|Amazon|3:21 AM|14
22July2016|Amazon|3:22 AM|14
22July2016|Amazon|3:23 AM|14

例如。 Apple 数据用于 3:22 ，因此请跳过并转到下一个。但同时要确保——在我的例子中，计数是 2（在本例中是 Apple+Amazon）10。 CE1 = 苹果，PE1 = 亚马逊早期输出：

3:20|31
3:21|30
3:23|28

需要新输出：

3:20|15|16
3:21|16|14
so on.

代码与CE1、CE2、PE2等匹配

谢谢

Answer 1

如果数据驻留在 MS Access 中，请考虑 crosstab query，您可以使用 PIVOT 子句显式指定 10 列：

TRANSFORM SUM(t.Price) As SumPrice
SELECT t.[Time]
FROM mytable t
GROUP BY t.[Time]
PIVOT t.ticker IN ('Apple', 'Cisco', 'Oracle', 'Microsoft', 'IBM',
                   'Amazon', 'Alphabet', 'Intel', 'Adobe', 'SAP')

或者，运行条件聚合，再次明确指定具有 10 个表达式的 10 列。这种方法允许灵活地重命名列，例如后缀 _Value.

SELECT t.Time,
       SUM(IIF(t.ticker = 'Apple', t.Price, NULL)) AS CE1_Value,
       SUM(IIF(t.ticker = 'Cisco', t.Price, NULL)) AS CE2_Value,
       SUM(IIF(t.ticker = 'Oracle', t.Price, NULL)) AS CE3_Value,
       ...
       SUM(IIF(t.ticker = 'Amazon', t.Price, NULL)) AS PE1_Value,
       SUM(IIF(t.ticker = 'Alphabet', t.Price, NULL)) AS PE2_Value,
       SUM(IIF(t.ticker = 'Intel', t.Price, NULL)) AS PE3_Value,
       ...
FROM mytable t
GROUP BY t.Time

是否需要将值传递到查询中，正如我在运行时在您的 consider ADO parameterization. Alternatively, with DAO (Access' default API) you can save a prepared statement as an Access query object using PARAMETERS clause and bind values in code using QueryDefs 中所建议的那样。这两种方法都避免了混乱的连接和标点符号，如 mysql 变量中所示。

DAO

SQL （另存为 Access 查询；无引号或连接）

PARAMETERS [DateIn] Date, [Qty_Param] Long,
           [CE1_Param] Text, [CE2_Param] Text, [CE3_Param] Text, [CE4_Param] Text, [CE5_Param] Text, 
           [PE1_Param] Text, [PE2_Param] Text, [PE3_Param] Text, [PE4_Param] Text, [PE5_Param] Text;
SELECT t.DateTr, t.[Time],
       [Qty_Param] * SUM(IIF(t.ticker = [CE1_Param], t.Price, NULL)) AS CE1_Value,
       [Qty_Param] * SUM(IIF(t.ticker = [CE2_Param], t.Price, NULL)) AS CE2_Value,
       [Qty_Param] * SUM(IIF(t.ticker = [CE3_Param], t.Price, NULL)) AS CE3_Value,
       ...
       [Qty_Param] * SUM(IIF(t.ticker = [PE1_Param], t.Price, NULL)) AS PE1_Value,
       [Qty_Param] * SUM(IIF(t.ticker = [PE2_Param], t.Price, NULL)) AS PE2_Value,
       [Qty_Param] * SUM(IIF(t.ticker = [PE3_Param], t.Price, NULL)) AS PE3_Value,
       ...
FROM mytable t
WHERE t.DateTr = [DateIn])
GROUP BY t.DateTr, t.[Time]

VBA

Dim db As DAO.Database, rst As DAO.Recordset, qdef As DAO.QueryDef

Set db = CurrentDb
Set qdef = db.QueryDefs("mySavedQuery")

With qdef
    .Parameters("DateIn") = CDate("2016-07-22")
    .Parameters("Qty_Param") = ###
    .Parameters("CE1_Param") = "Apple"
    .Parameters("CE2_Param") = "..."
    .Parameters("CE3_Param") = "..."
    ...
    .Parameters("PE1_Param") = "Amazon"
    .Parameters("PE2_Param") = "..."
    .Parameters("PE3_Param") = "..."
    ...
End With

Set rst = qdef.OpenRecordset()

...

rst.Close: qdef.Close
Set rst = Nothing: Set qdef = Nothing: Set db = Nothing

Answer 2

下面是我最终的工作 SQL 代码。

SELECT Data.Time, 
       Round(Sum(IIf(Data.Ticker='Apple',Data.price,Null)),2) AS CE1_Value, 
       Round(Sum(IIf(Data.Ticker='Amazon',Data.price,Null)),2) AS CE2_Value, 
       Round(Sum(IIf(Data.Ticker='Intel',Data.price,Null)),2) AS PE1_Value, 
       Round(Sum(IIf(Data.ticker='Microsoft',Data.price,Null)),2) AS PE2_Value 
FROM Data 
WHERE (((Data.Time)>#12/30/1899 12:20:0#) AND ((Data.DateTr)=#6/18/2019#)) 
GROUP BY Data.Time, Data.DateTr 
HAVING (((Sum(IIf(Data.Ticker='Apple',Data.price,Null))) Is Not Null) 
   AND  ((Sum(IIf(Data.Ticker='Amazon',Data.price,Null))) Is Not Null) 
   AND  ((Sum(IIf(Data.Ticker='Intel',Data.price,Null))) Is Not Null) 
   AND  ((Sum(IIf(Data.ticker='Microsoft',Data.price,Null))) Is Not Null)) 
ORDER BY Data.Time;

如何在 SQL 上使用 PIVOT 以及列数检查和分组等附加条件？

How to use PIVOT on SQL with additional criteria like colum count check and grouping?

sql

vbscript

ms-access

pivot

DAO