Select 每个分组列只有 2 行数据

Question

I have three tables in my database, for the purposes of discussion let's say they are:

USERS   
-----
id
display_name

ROLES
-----
id
role

USER_ROLES
----------
user_role_id
user_id
role_id

As you can see 1 user can have multiple roles. Now I have successfully grouped them as their roles with the following query.

SELECT r.`name`, GROUP_CONCAT(u.display_name) AS users_for_role
FROM users u,
  roles r,
  user_roles ur
WHERE u.id = ur.user_id
  AND r.id = ur.role_id
GROUP BY r.id;

And got the result like

| ROLE_USER         | user1,user6,user4 |
| ROLE_CELEBRITY    | user4,user5,user6 |
| ROLE_MUSICIAN     | user4             |

Now what I'm trying to achieve is I want to limit the select count to max 2. So my expected result would be ROLE_USER selects only "user1" and "user6" MAX

Answer 1

One trick you could use here would be to just wrap your current call to GROUP_CONCAT with SUBSTRING_INDEX , to retain only at most the first two entries:

SELECT r.name, SUBSTRING_INDEX(GROUP_CONCAT(u.display_name), ',', 2) AS users_for_role
FROM users u
INNER JOIN roles r ON u.id = ur.user_id
INNER JOIN user_roles ur ON r.id = ur.role_id
GROUP BY r.id;

Another more formal approach would be to use a subquery with ROW_NUMBER to restrict to the first two entries per name:

WITH cte AS (
    SELECT r.id, r.name, u.display_name,
           ROW_NUMBER() OVER (PARTITION BY r.name ORDER BY u.display_name) rn
    FROM users u
    INNER JOIN roles r ON u.id = ur.user_id
    INNER JOIN user_roles ur ON r.id = ur.role_id
)

SELECT name, GROUP_CONCAT(display_name) AS users_for_role
FROM cte
WHERE rn <= 2
GROUP BY id;