I create a view with following codes
SELECT
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T1' AS sno,
YEAR(okuma_tarihi) AS Yillar,
SUM(toplam_kullanim_T1) AS TotalUsageValue, 'T1' AS UsageType
FROM
TblSayacOkumalari
GROUP BY
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T1', YEAR(okuma_tarihi)
UNION ALL
SELECT
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T2' AS sno,
YEAR(okuma_tarihi) AS Yillar,
SUM(toplam_kullanim_T2) AS TotalUsageValue, 'T2' AS UsageType
FROM
TblSayacOkumalari
GROUP BY
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T1', YEAR(okuma_tarihi)
UNION ALL
SELECT
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T3' AS sno,
YEAR(okuma_tarihi) AS Yillar,
SUM(toplam_kullanim_T3) AS TotalUsageValue, 'T3' AS UsageType
FROM
TblSayacOkumalari
GROUP BY
CONVERT(NVARCHAR, YEAR(okuma_tarihi)) + 'T1', YEAR(okuma_tarihi)
I want to define CONVERT(nvarchar, YEAR(okuma_tarihi)) + 'T1' AS sno
as a primary key is that possible? If is this possible how can I do?
You cannot create a primary key on a view. In SQL Server you can create an index on a view but that is different to creating a primary key.
If you give us more information as to why you want a key on your view, perhaps we can help with that.
You may not be able to create a primary key (per say) but if your view is based on a table with a primary key and the key is included in the view, then the primary key will be reflected in the view also. Applications requiring a primary key may accept the view as it is the case with Lightswitch.
A little late to this party - but this also works well:
CREATE VIEW [ABC].[View_SomeDataUniqueKey]
AS
SELECT
CAST(CONCAT(CAST([ID] AS VARCHAR(4)),
CAST(ROW_NUMBER() OVER(ORDER BY [ID] ASC) as VARCHAR(4))
) AS int) AS [UniqueId]
,[ID]
FROM SOME_TABLE JOIN SOME_OTHER_TABLE
GO
In my case the join resulted in [ID] - the primary key being repeated up to 5 times (associated different unique data)
The nice trick with this is that the original ID can be determined from each UniqueID
effectively [ID]+RowNumber() = 11, 12, 13, 14, 21, 22, 23, 24 etc.
If you add RowNumber() and [ID] back into the view - you can easily determine your original key from the data.
But - this is not something that should be committed to a table because I am fairly sure that the RowNumber() of a view will never be reliably the same as the underlying data alters, even with the OVER(ORDER BY [ID] ASC) to try and help it.
Example output ( Select UniqueId, ID, ROWNR, Name from [REF].[View_Systems] ) :
UniqueId ID ROWNR Name
11 1 1 Amazon A
12 1 2 Amazon B
13 1 3 Amazon C
14 1 4 Amazon D
15 1 5 Amazon E
Table1:
[ID] [Name]
1 Amazon
Table2:
[ID] [Version]
1 A
1 B
1 C
1 D
1 E
CREATE VIEW [REF].[View_Systems]
AS
SELECT
CAST(CONCAT(CAST(TABA.[ID] AS VARCHAR(4)),
CAST(ROW_NUMBER() OVER(ORDER BY TABA.[ID] ASC) as VARCHAR(4))
) AS int) AS [UniqueId]
,TABA.[ID]
,ROW_NUMBER() OVER(ORDER BY TABA.[ID] ASC) AS ROWNR
,TABA.[Name]
FROM [Ref].[Table1] TABA LEFT JOIN [Ref].[Table2] TABB ON TABA.[ID] = TABB.[ID]
GO
This worked for me..
select ROW_NUMBER() over (order by column_name_of your choice ) as pri_key, the other columns of the view
I got the error "The table/view 'dbo.vMyView' does not have a primary key defined" after I created a view in SQL server query designer. I solved the problem by using ISNULL on a column to force entity framework to use it as a primary key. You might have to restart visual studio to get the warnings to go away.
CREATE VIEW [dbo].[vMyView]
AS
SELECT ISNULL(Id, -1) AS IdPrimaryKey, Name
FROM dbo.MyTable