One to many join with group by - sql

Question

I have two tables. one table is named Shopper and it looks like
SHOPPER_ID | SHOPPER_NAME |
-------------------------
1 | Marianna |
2 | Jason |
and another table named Order has information like Date on the order
ORDER_ID | SHOPPER_ID | DATE
----------------------------------
1 | 1 | 08/09/2012
2 | 1 | 08/08/2012
Now I want to do a query that joins two tables and group by SHOPPER_ID, because one shopper can have multiple orders, I want to pick the latest order base on DATE value.
My query looks like:
Select * from Shopper as s join Order as o
on s.SHOPPER_ID = o.SHOPPER_ID
group by s.SHOPPER_ID
The query is wrong right now because I don't know how to apply the filter to only get the latest order. Thanks in advance!

Answer 1

I suggest using a sub-select:
Select s.SHOPPER_ID, s.SHOPPER_NAME, o.MAX_DATE
from Shopper s
INNER join (SELECT SHOPPER_ID, MAX(DATE) AS MAX_DATE
FROM ORDER
GROUP BY SHOPPER_ID) o
on s.SHOPPER_ID = o.SHOPPER_ID
Best of luck.

Answer 2

Easy way is use row_number to find the lastest order
SQL Fiddle Demo
SELECT *
FROM
(SELECT S.*,
O.[ORDER_ID], O.[DATE],
ROW_NUMBER() OVER ( PARTITION BY S.SHOPPER_ID
ORDER BY [DATE] DESC) as rn
FROM Shopper S
JOIN Orders O
ON S.SHOPPER_ID = O.SHOPPER_ID
) T
WHERE rn = 1

Answer 3

SELECT *
FROM
Shopper s
CROSS APPLY
(
SELECT TOP 1 *
FROM
Order o
WHERE
s.SHOPPER_ID = o.SHOPPER_ID
ORDER BY
o.DATE DESC
) o;

Answer 4

You need a subquery to get the last order per shopper, and then join that with the shopper and order tables to get the name of the shopper and the order id
SELECT ss.SHOPPER_ID, ss.SHOPPER_NAME, oo.ORDER_ID LAST_ORDER
FROM (SELECT o.SHOPPER_ID, MAX(o.DATE) [DATE]
FROM Shopper s
INNER JOIN Order o
ON s.SHOPPER_ID = o.SHOPPER_ID
GROUP BY o.SHOPPER_ID) mo
INNER JOIN Shopper ss
ON mo.SHOPPER_ID = ss.SHOPPER_ID
INNER JOIN Order oo
ON mo.SHOPPER_ID = oo.SHOPPER_ID AND mo.DATE = oo.DATE
Here's the SQL Fiddle to try it out

Answer 5

Select s.*, o1.*
From Order as o1
left join Order as o2
on (o1.SHOPPER_ID = o2.SHOPPER_ID and o1.DATE < o2.DATE)
join Shopper as s
on (s.SHOPPER_ID = o1.SHOPPER_ID )
where o2.DATE is NULL;
Join Order table to itself, looking for newer Orders to join it to. The "left" join means that every row in the Order table will be kept in the results even if it cannot be joined to a newer order for that customer.
The "where" discards all of the rows where a newer order was found. This leaves you only with only the most recent Orders.
Join those results to the Shopper table to include the shopper data.
Edit: I suggested this answer because JOINs are much faster for a Database than sub-selects.