Question

I have a stored procedure used for updating records in a SQL Server database.

I want to use the existing value within the database if a match is found and a null is provided as the variable.

Often only partial data is available so the procedure is executed at several stages to complete the record. I do not always have all variables at each execution.

Why is the below CASE not working? If a null value is provided, it replaces existing data within the field.

ALTER PROCEDURE [dbo].[sp_UPSERT_Test] 

@ID int,
@Account_ID varchar(15)
@Name varchar(15)

AS 
BEGIN 

MERGE INTO dbo.Database_Log AS target  
USING ( 
   SELECT @ID, @Account_ID ,@Account_ID)  
   AS source (ID, Account_ID, Name)
ON target.ID = source.ID 

WHEN MATCHED
THEN UPDATE
SET
    Account_ID = CASE 
    WHEN source.Account_ID = NULL 
    THEN target.Account_ID 
    ELSE source.Account_ID 
        END
    ,Name = CASE 
    WHEN source.Name = NULL 
    THEN target.Name 
    ELSE source.Name 
        END)

WHEN NOT MATCHED  
THEN INSERT (Account_ID, Name)  
       VALUES (Account_ID, Name);

RETURN
END

Answer 1

Maybe because you’re not properly checking for NULL ? NULL is not a value you can compare with the equality operator – you must use IS NULL

Try this:

 SET      
    Account_ID = CASE 
                    WHEN source.Account_ID IS NULL 
                    THEN target.Account_ID 
                    ELSE source.Account_ID 
                 END

Use IS NULL instead of = NULL