Question

我想知道给定表中每个记录的关联数量。其中一些关联附加了一些条件

到目前为止我已经

了

-- Count app associations
SELECT 
distinct a.name, 
COALESCE(v.count, 0) as visitors, 
COALESCE(am.count, 0) AS auto_messages,
COALESCE(c.count, 0) AS conversations
FROM apps a
LEFT JOIN (SELECT app_id, count(*) AS count FROM visitors GROUP BY 1) v ON a.id = v.app_id
LEFT JOIN (SELECT app_id, count(*) AS count FROM auto_messages GROUP BY 1) am ON a.id = am.app_id
LEFT JOIN (
    SELECT DISTINCT c.id, app_id, count(c) AS count 
    FROM conversations c LEFT JOIN messages m ON m.conversation_id = c.id
    WHERE m.visitor_id IS NOT NULL
    GROUP BY c.id) c ON a.id = c.app_id
WHERE a.test = false
ORDER BY visitors DESC;

我遇到了对话的最后一个连接语句的问题。我想计算至少有一条消息的会话数，其中visitor_id不为null。出于某种原因，我为每个应用程序获得了多条记录，即。对话没有正确分组。

有什么想法吗？

Answer 1

我的直觉，基于对大图的有限理解：在conversations选择的嵌套查询中，

删除DISTINCT
从SELECT list

c.id

GROUP BY c.app_id而不是c.id

编辑：试试这个

...
LEFT JOIN (
    SELECT app_id, count(*) AS count 
    FROM conversations c1
    WHERE 
    EXISTS (
        SELECT *
        FROM messages m
        WHERE m.conversation_id = c1.id and
            M.visitor_id IS NOT NULL
    )
    GROUP BY c1.app_id) c
ON a.id = c.app_id

计算来自多个表的关联

1 个答案: