views:

39

answers:

1

Hey all,

I've been roasting my brain with my limited SQL knowledge while attempting to come up with a query to run a statistic on my orders database.

Table ORDERS is laid out like this:

CustomerID ProductID (etc)
1              10
1              10
1              11
2              10
4              9

Each purchase is recorded with the customer id and the product ID - there CAN be multiple records for the same customer, and even multiple records with the same customer and product.

I need to come up with a query that can return the amount of customers who bought between X and X distinct products - for example, 3 customers bought less then 5 different products, 10 bought from 5-10 different products, 1 bought over 10 different products.

I'm pretty sure this has something to do with derived tables, but advanced SQL is a new fairly craft to me. Any help would be appreciated!

+2  A: 

Try this:

SELECT T1.products_bought, COUNT(T2.cnt) AS total
FROM (
    SELECT '<5' AS products_bought, 0 AS a, 4 AS b
    UNION ALL
    SELECT '5-10', 5, 10
    UNION ALL
    SELECT '>10', 11, 999999
) T1
LEFT JOIN
(
    SELECT COUNT(DISTINCT ProductID) AS cnt
    FROM ORDERS
    GROUP BY CustomerID
) T2
ON T2.cnt BETWEEN T1.a AND T1.b
GROUP BY a, b

Result:

products_bought  total
<5               3    
5-10             0    
>10              0    
Mark Byers
Works perfectly Mark, you're a lifesaver.
MarathonStudios