我在Tableau中插入SQL查询时出现此错误消息。如何解决这个问题?
SQL Server当前是否不允许在子查询中使用CTE?
下面是我在
中插入查询时的Tableau输出[Microsoft] [SQL Server Native Client 11.0] [SQL Server]关键字“WITH”附近的语法不正确。
[Microsoft] [SQL Server Native Client 11.0] [SQL Server]关键字“with”附近的语法不正确。如果此语句是公用表 表达式,xmlnamespaces子句或更改跟踪上下文 子句,前一个语句必须以分号结束。
[Microsoft] [SQL Server Native Client 11.0] [SQL Server]')'附近的语法不正确。
以下是我当前的CTE查询(递归)
WITH shiftHours AS (
SELECT RowID,
y.EMPLOYEENAME AS EMPLOYEENAME,
-- flatten the first hour to remove the minutes and get the initial current hour
DATEADD(hour, DATEDIFF(hour, 0, ShiftA_Start), 0) AS currentHour,
ShiftA_Start,
ShiftA_End,
DATEPART(hour, ShiftA_Start) AS hourOrdinal,
-- determine how much of the first hour is applicable. if it is minute 0 then the whole hour counts
CAST(CASE
WHEN DATEADD(hour, DATEDIFF(hour, 0, ShiftA_Start), 0) = DATEADD(hour, DATEDIFF(hour, 0, ShiftA_End), 0) THEN DATEDIFF(minute, ShiftA_Start, ShiftA_End) / 60.0
WHEN DATEPART(minute, ShiftA_Start) = 0 THEN 1.0
ELSE (60 - DATEPART(minute, ShiftA_Start)) / 60.0
END AS DECIMAL(5,3)) AS hourValue
FROM (
-- use a ROW_NUMBER() to generate row IDs for the shifts to ensure each row is unique once it gets to the pivot
SELECT ROW_NUMBER() OVER(ORDER BY ShiftA_Start, ShiftA_End) AS RowID,
EMPLOYEENAME,
ShiftA_Start,
ShiftA_End
FROM (
-- this is where the data gets pulled from the source table and where the data types are converted from string to DATETIME
SELECT
EMPLOYEENAME,
CONVERT(DATETIME, LEFT(SHIFTA_start, 17), 103) AS ShiftA_Start,
CONVERT(DATETIME, LEFT(SHIFTA_end, 17), 103) AS ShiftA_End
from [TableName].[dbo].[TMS_People]
where
CONVERT(DATETIME, LEFT(SHIFTA_START, 17), 103) IS NOT NULL AND CONVERT(DATETIME, LEFT(SHIFTA_END, 17), 103) IS NOT NULL
AND CONVERT(DATETIME, LEFT(SHIFTA_START, 17), 103) != CONVERT(DATETIME, LEFT(SHIFTA_END, 17), 103)
-- this is also where you would add any filtering from the source table such as date ranges
) x
) AS y
UNION ALL
SELECT RowID,
EMPLOYEENAME,
-- add an hour to the currentHour each time the recursive CTE is called
DATEADD(hour, 1, currentHour) AS currentHour,
ShiftA_Start,
ShiftA_End,
DATEPART(hour, DATEADD(hour, 1, currentHour)) AS hourOrdinal,
CAST(CASE
-- when this is the last time period determine the amount of the hour that is applicable
WHEN DATEADD(hour, 2, currentHour) > ShiftA_End THEN DATEPART(minute, ShiftA_End) / 60.0
ELSE 1
END AS DECIMAL(5,3)) AS hourValue
from shiftHours
-- contine recursion until the next hour is after the ShiftEnd
WHERE DATEADD(hour, 1, currentHour) < ShiftA_End
)
SELECT *
FROM (
SELECT RowID,
EMPLOYEENAME,
ShiftA_Start,
ShiftA_End,
hourValue,
hourOrdinal
from shiftHours
) AS t
PIVOT (
SUM(hourValue)
FOR hourOrdinal IN ([0], [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21], [22], [23])
) AS pvt
OPTION (MAXRECURSION 0);
答案 0 :(得分:2)
Tableau不支持SQL Server CTE功能。
您应该尝试使用视图或子查询或存储过程/表值函数来实现相同的逻辑。
请在下面抛出线程CTE(SQL Server)不在Tableau中运行。
https://community.tableau.com/thread/105965。
查看示例代码。对于CTE
如果我们有用户表(userId,userName,managerId)
CREATE VIEW vUSER AS
WITH UserCTE AS (
SELECT userId, userName, managerId,0 AS steps
FROM dbo.Users
WHERE userId = null
UNION ALL
SELECT mgr.userId, mgr.userName, mgr.managerId, usr.steps +1 AS steps
FROM UserCTE AS usr
INNER JOIN dbo.Users AS mgr
ON usr.managerId = mgr.userId
)
SELECT * FROM UserCTE ;
创建一个sp即MyProc并将所有查询放入其中 即
CREATE procedure dbo.MyProc AS
WITH UserCTE AS (
SELECT userId, userName, managerId,0 AS steps
FROM dbo.Users
WHERE userId = null
UNION ALL
SELECT mgr.userId, mgr.userName, mgr.managerId, usr.steps +1 AS steps
FROM UserCTE AS usr
INNER JOIN dbo.Users AS mgr
ON usr.managerId = mgr.userId
)
SELECT * FROM UserCTE ;
2使用sp_addlinkedserver创建一个链接服务器,即我告诉的是Local
3使用openquery在视图中调用sp。
CREATE VIEW dbo.MyView
AS(
SELECT *
FROM openquery(Local,'exec mydb.dbo.MyProc'))
http://capnjosh.com/blog/how-to-call-a-stored-procedure-from-a-view-in-sql-server/