Question

I have a table that looks like this:-

TABLEA

SERIAL  SKU GRADE   ID
HA501   R2022   2   2011063
HA501   R2022   1   2011052
HA502   R2033   2   2011051
HA502   R2033   3   2011048
HA503   R2044   1   2011034
HA503   R2044   2   2011023

I wish to extract the above SERIAL, SKU & GRADE based on the latest ID, like this:

FINAL OUTPUT            
SERIAL  SKU GRADE   ID
HA501   R2022   2   2011063
HA502   R2033   2   2011051
HA503   R2044   1   2011034

I was trying to using SQL:

select 
     SERIAL, SKU, GRADE, MAX(ID) 
from tableA
group by SERIAL, SKU, GRADE

but somehow the output doesn’t seems right.

Can someone help me ?

Thanks.

Answer 1

Should you be using SQL Server 2005 and newer, you could use a CTE with the ROW_NUMBER construct:

;WITH DataForLastID AS
(
    SELECT 
        SERIAL, SKU, GRADE, ID,
        ROW_NUMBER() OVER(PARTITION BY SERIAL, SKU ORDER BY ID DESC) AS 'RowNum'
    FROM
       dbo.TABLEA
)
SELECT
    SERIAL, SKU, GRADE, ID
FROM
   DataForLastID
WHERE
    RowNum = 1

This will “partition” your data by Serial,SKU,Grade and for each group of the same values, it will sequentially number the rows starting at 1, ordered by descending ID (so the largest/newest ID will be RowNum = 1).