How to “Implode” (de-normalize/concat) multiple columns into a single column?

后端 未结 4 1944
别跟我提以往
别跟我提以往 2021-01-17 23:46

I have a query which outputs something like this:

+-------+----+--------------+
| F_KEY | EV | OTHER_COLUMN |
+-------+----+--------------+
| 100   | 1  | ..         


        
4条回答
  •  不知归路
    2021-01-18 00:23

    You can do it easily enough with a function: Assumes you are going to be searching the same column/table all the time. Dynamic SQL needed if you want to be able to vary the columns/tables

    CREATE FUNCTION [dbo].[fn_recursion]
    (@F_KEY int)  
    RETURNS varchar(2000) AS 
    BEGIN 
    
        DECLARE @ReturnVal Varchar(2000)
    
        SELECT @ReturnVal = COALESCE(@ReturnVal + ', ', '') + EV
        FROM TABLE2 
        WHERE @F_KEY = @F_KEY
    
        RETURN ISNULL(@ReturnVal,'')
    
    END
    GO
    
    
    SELECT
        F_KEY,
        EV = [dbo].[fn_recursion](F_KEY),
        OTHER_COLUMN
    FROM
        TABLE1
    JOIN
        TABLE2 ON F_KEY = TABLE2.ID
    WHERE
        EVENT_TIME BETWEEN '2011-01-01 00:00:00.000' AND '2011-12-31 23:59:59.999'
    ORDER BY
        EVENT_TIME ASC
    
    GO
    
    DROP FUNCTION [dbo].[fn_recursion]
    GO
    

提交回复
热议问题