我可能不会这样做，因为它比前面提到的CASE结构效率更低——除非您为两个查询都有覆盖索引。不管怎样，对于类似的问题，这都是一个有用的技巧:

SELECT OrderId, MAX(Price) as Price FROM (
   SELECT o.OrderId, o.NegotiatedPrice as Price FROM Order o
   UNION ALL
   SELECT o.OrderId, o.SuggestedPrice as Price FROM Order o
) as A
GROUP BY OrderId

在MemSQL中执行以下操作:

-- DROP FUNCTION IF EXISTS InlineMax;
DELIMITER //
CREATE FUNCTION InlineMax(val1 INT, val2 INT) RETURNS INT AS
DECLARE
  val3 INT = 0;
BEGIN
 IF val1 > val2 THEN
   RETURN val1;
 ELSE
   RETURN val2;
 END IF; 
END //
DELIMITER ;

SELECT InlineMax(1,2) as test;

CREATE FUNCTION [dbo].[fnMax] (@p1 INT, @p2 INT)
RETURNS INT
AS BEGIN

    DECLARE @Result INT

    SET @p2 = COALESCE(@p2, @p1)

    SELECT
        @Result = (
                   SELECT
                    CASE WHEN @p1 > @p2 THEN @p1
                         ELSE @p2
                    END
                  )

    RETURN @Result

END

SELECT o.OrderId,   
--MAX(o.NegotiatedPrice, o.SuggestedPrice)  
(SELECT MAX(v) FROM (VALUES (o.NegotiatedPrice), (o.SuggestedPrice)) AS value(v)) as ChoosenPrice  
FROM Order o

这里有一个案例示例，应该处理null，并将与旧版本的MSSQL一起工作。这是基于一个流行例子中的内联函数:

case
  when a >= b then a
  else isnull(b,a)
end

子查询可以从Outer查询访问列，因此可以使用这种方法跨列使用MAX等聚合。(不过，当涉及到大量列时可能更有用)

;WITH [Order] AS
(
SELECT 1 AS OrderId, 100 AS NegotiatedPrice, 110 AS SuggestedPrice UNION ALL
SELECT 2 AS OrderId, 1000 AS NegotiatedPrice, 50 AS SuggestedPrice
)
SELECT
       o.OrderId, 
       (SELECT MAX(price)FROM 
           (SELECT o.NegotiatedPrice AS price 
            UNION ALL SELECT o.SuggestedPrice) d) 
        AS MaxPrice 
FROM  [Order]  o