Question

我有多对多的帐户＆lt; - ＆gt; 用户关系。我正在尝试提取我和其他用户共享的所有帐户。到目前为止，我有一些我从this回答：

修改的内容

SELECT * 
    FROM user
    JOIN account_user ON account_user.user_id = user.id
    JOIN account ON account.id = account_user.account_id
   WHERE user.id IN ({my_id},{other_user_id})
GROUP BY account.id
  HAVING COUNT(DISTINCT user.id) = 2

这似乎是在正确的轨道上，但account_user表有一个'角色'字段，我想确保我得到的记录包含其他用户的角色，而不是我的。我如何调整这个以实现这一目标？

Answer 1

在我的示例中，'user2'是另一个用户，'user1'是你：

-- query should work in most RDBMS
SELECT A1.account,
       AU1.role AS user_2_account_role,
       U1.uname AS user_2_name
FROM @account AS A1
JOIN @account_users AS AU1 ON AU1.account_id = A1.id
JOIN @user AS U1 ON U1.id = AU1.user_id
WHERE EXISTS(
    -- get accounts that are 'mine'
    SELECT *
    FROM @user AS U2
    JOIN @account_users AS AU2 ON AU2.user_id = U2.id
    WHERE U2.uname = 'user1' -- 'mine'
    AND AU2.account_id = AU1.account_id
)
AND U1.uname = 'user2' -- other user

我在SQL Server中准备的示例数据：

-- sample data in SQL SERVER
DECLARE @user TABLE(uname VARCHAR(20), id INT);
INSERT INTO @user(uname, id) 
VALUES('user1', 1),
      ('user2', 2),
      ('user3', 3),
      ('user4', 4);

DECLARE @account TABLE(account VARCHAR(20), id INT);
INSERT INTO @account(account, id)
VALUES('account1', 1),
      ('account2', 2),
      ('account3', 3),
      ('account4', 4);

DECLARE @account_users TABLE(account_id INT, user_id INT, role VARCHAR(10));
INSERT INTO @account_users(account_id, user_id, role)
VALUES(1, 1, 'user1_role'),
      (1, 2, 'user2_role'),
      (2, 1, 'user1_role'),
      (3, 3, 'user3_role'),
      (3, 4, 'user4_role');

Answer 2

这在我的Oracle下工作，同样的基本思想应该在任何DBMS下工作：

SELECT account.*, role
FROM account
    JOIN account_user ON account_id = account.id AND user_id = :other_user_id
WHERE
    account.id IN (SELECT account_id FROM account_user WHERE user_id = :my_id)

用简单的英语：

为其他用户加入account和account_user。
仅接受那些也连接到我的用户的帐户。

如果需要，您可以轻松加入user（如您的SQL所示，但不是您的问题）。

精炼双倍多对多JOIN

2 个答案: