如何在 MYSQL 中使用 COUNT 和 GROUP BY 同时仍显示所有单独的行？

Question

I am just trying some practise exercises with MYSQL. I have a dataset where I would like to get the names of customers who ordered two or more different kinds of item and how many of each kind of item they bought.

The query below gives me a row for each name of the purchaser. However, I also want to display what types of items they bought and how many of them. Ideally I would like to have the same number of rows for each customer for how many different items they bought.

SELECT firstname, familyname, description, quantity
FROM customers c
JOIN orders o ON o.custID = c.custID
JOIN lineitems l on o.orderID = l.orderID
JOIN items i on l.itemID = i.itemID
GROUP BY firstname
HAVING count(description) 

The query below does give me a row for each item, how many items that person bought, and the name of the purchaser. However, it does not filter for customers who only bought one specific item anymore.

SELECT firstname, familyname, description, quantity
FROM customers c
JOIN orders o ON o.custID = c.custID
JOIN lineitems l on o.orderID = l.orderID
JOIN items i on l.itemID = i.itemID
WHERE EXISTS(
    SELECT *
    FROM customers
    GROUP BY firstname
    HAVING count(description) >= 2)

Basically I would like to combine both approaches where there are multiple rows for specific item for each customer, while also filtering out customers who only bought one type of item.

Answer 1

If you are running MySQL 8.0, you can do a window count in a subquery and filter in the outer query, like:

SELECT *
FROM (
    SELECT 
        c.firstname, 
        c.familyname, 
        i.description, 
        l.quantity, 
        COUNT(*) OVER(PARTITION BY c.custID) cnt
    FROM customers c
    JOIN orders o ON o.custID = c.custID
    JOIN lineitems l on o.orderID = l.orderID
    JOIN items i on l.itemID = i.itemID
) x
WHERE cnt > 1

Note: it is a good practice to prefix column names with the alias of the table they belong to; this makes the query more readable and avoid clashes when the same column name exists across tables. I made a few assumptions and updated the query accordingly.