如何基于SQL Server的另一列中的最高价值为每个客户获得一行

Question

I have a table that contains some customer information:

--------------------------------------------
|CustID  CustType  OrderType  CountofOrders|
|------------------------------------------|
|  1        C         DD          23       |
|  2       IC         DE          10       |
|  2       IC         DR           7       |
|  2       IC         DS           7       |
|  3        C         DR          14       |
|  3        C         DS          19       |
|  4       IC         DS           5       |
|  4       IC         DR           5       |
|  4       IC         DE           3       |
|  5       IC         DR           7       |
|  5       IC         DS           7       |
|  5       IC         DE           7       |
--------------------------------------------

I want to pull all records for Customers when CustType is not 'IC', regardless of how many rows they have.

So for instance, CustID 1 would return 1 row while CustID 3 would return 2 rows. When the CustType is 'IC', I only want to pull a single row for that customer based on the highest CountofOrders.

So for CustID 2, it should pull the record with OrderType 'DE' and CountofOrders as 10.

If the 'IC' customer has the same count for the various OrderTypes, then I want to get the row where the OrderType is 'DE'. If the OrderType is not 'DE', then I want to get the row with the OrderType as 'DS'. In short, priority should be given to OrderType 'DE', followed by 'DS', 'DR' and 'DD'

So my final output should look something like this:

-------------------------------------------
|CustID  CustType  OrderType  CountofOrders|
-------------------------------------------|
|  1        C         DD           23      |
|  2       IC         DE           10      |
|  3        C         DR           14      |
|  3        C         DS           19      |
|  4       IC         DS            5      |
|  5       IC         DE            7      |
--------------------------------------------

I was able to use the following query to get the highest for when the CountofOrders is not the same, but I'm having a hard time figuring out how to get a row based on the OrderType when the CountofOrders is the same between the different OrderTypes.

SELECT 
    a.CustID, a.CustType, a.OrderType, a.CountofOrders
FROM 
    CustInfo a 
WHERE 
    a.CountofOrders = (SELECT MAX(CountofOrders) 
                       FROM CustInfo b 
                       WHERE a.CustID = b.CustID)

Any help would be greatly appreciated. Thanks.

Answer 1

You can do this using ROW_NUMBER pretty easily. This returns the results you stated you wanted from your sample data.

DECLARE @Customer table
(
    CustID int
    , CustType varchar(10)
    , OrderType char(2)
    , CountofOrders int
)

INSERT @Customer
(
    CustID
    , CustType
    , OrderType
    , CountofOrders
)
VALUES
(1, 'C', 'DD', 23)
, (2, 'IC', 'DE', 10)
, (2, 'IC', 'DR', 7)
, (2, 'IC', 'DS', 7)
, (3, 'C', 'DR', 14)
, (3, 'C', 'DS', 19)
, (4, 'IC', 'DS', 5)
, (4, 'IC', 'DR', 5)
, (4, 'IC', 'DE', 3)
, (5, 'IC', 'DR', 7)
, (5, 'IC', 'DS', 7)
, (5, 'IC', 'DE', 7)

SELECT x.CustID
    , x.CustType
    , x.OrderType
    , x.CountofOrders
FROM
(
    SELECT *
         , ROW_NUMBER() over(partition by CustID order by CountOfOrders desc
            , case OrderType 
                when 'DE' then 1
                when 'DS' then 2
                when 'DR' then 3
                when 'DD' then 4
                else 5
            end) as RowNum
    FROM @Customer
) x
WHERE x.CustType <> 'IC'
    OR x.RowNum = 1
ORDER BY CustID
    , CountofOrders

Answer 2

I think your requirement cannot be done in query level because the order is based on priority. I mean you have to use something like stored procedure or do the ordering process in your application