Question

有2个表：mess_chats，mess_posts。我需要将它们合并到chatID列，然后按时间戳排序，以便最新的帖子位于顶部，最后按chatID分组。

这是我的代码：

SELECT * FROM (
  SELECT   mess_chats.*, mess_posts.*
  FROM     mess_chats INNER JOIN mess_posts
           ON mess_chats.chatID = mess_posts.chatID
  WHERE    mess_chats.membersID="1"
  ORDER BY mess_posts.timestamp DESC
) AS tt
GROUP BY tt.chatID

这是错误：

#1060 - Duplicate column name 'chatID'

编辑：我找到了解决方案，首先我将mess_posts.chatID列名更改为mess_posts.pchatID，然后将sql代码更改为：

SELECT * FROM (
  SELECT   mess_chats.*, mess_posts.*
  FROM     mess_chats JOIN mess_posts
           ON mess_chats.chatID = mess_posts.pchatID
  WHERE    mess_chats.membersID="1"
  ORDER BY mess_posts.timestamp DESC
) AS tt
GROUP BY tt.pchatID

Answer 1

您在分组之前不需要订单，使用分组返回的结果是不确定的，这不能保证您为该帖子提供最近的聊天，但这是您可以的方式这样做，在chatid上使用mess_chats表的自联接加上最大时间戳

SELECT 
    c.*,
    p.* 
  FROM
    mess_chats c
    INNER JOIN 
    (SELECT chatID, MAX(`timestamp`) `timestamp` 
       FROM mess_chats GROUP BY chatID ) mc
      ON(c.chatID = mc.chatID AND c.`timestamp` = mc.`timestamp`)
    INNER JOIN mess_posts p 
      ON c.chatID = p.chatID 
  WHERE c.membersID = "1" 
  ORDER BY p.timestamp DESC

对于查询中的错误，我会说你在两个表中都有chatID列具有相同的名称，所以当你进行子选择时，你可以参考tt.chatID然后在结果中有两列chatID所以你看到重复的列名称错误，为此，你需要为列中的一个提供不同的别名

Answer 2

试试这个

SELECT * FROM (
    SELECT   mess_chats.*, mess_posts.*, mess_posts.chatID as 'post_chatID'
    FROM     mess_chats INNER JOIN mess_posts
        ON mess_chats.chatID = mess_posts.chatID
    WHERE    mess_chats.membersID="1"
    ORDER BY mess_posts.timestamp DESC
) AS tt
GROUP BY tt.post_chatID

在分组之前排序

2 个答案: