Question

我收到此错误，查询在SQLite中有效。

ActiveRecord::StatementInvalid (PG::GroupingError: ERROR:  COLUMN "receipts.is_read" must appear IN the GROUP BY clause OR be used IN an aggregate FUNCTION

不确定该怎么做，当我移除合并部件时也会发生这种情况。我不知道还有什么我可以指定它来使它工作。有什么想法吗？

SELECT DISTINCT conversations.id, COALESCE(CASE WHEN receipts.is_read = 't' THEN 't' END, 'f') AS READ, conversations.updated_at, p1.nickname, p2.nickname
FROM conversations
INNER JOIN notifications ON notifications.conversation_id = conversations.id
INNER JOIN receipts ON receipts.notification_id = notifications.id
INNER JOIN profiles p1 ON p1.id = notifications.sender_id
INNER JOIN profiles p2 ON p2.id = receipts.receiver_id
WHERE (receipts.receiver_id = #{SELF.id} ) AND notifications.sender_id != #{SELF.id} OR (notifications.sender_id = #{SELF.id}) AND receipts.receiver_id != #{SELF.id}
GROUP BY conversations.id
ORDER BY conversations.updated_at DESC

Answer 1

您的DISTINCT和GROUP BY以及一些未分组的列混合在一起。在用DISTINCT ON（以及其他一些清理工具）替换这些混乱之后，这可行，但它不清楚，你真正想要实现的目标：

SELECT DISTINCT ON (c.id)
       c.id, COALESCE(r.is_read, FALSE) AS read
     , c.updated_at, p1.nickname AS nickname1, p2.nickname AS nickname2
FROM   conversations c
JOIN   notifications n USING (conversation_id)
JOIN   receipts      r USING (notification_id)
JOIN   profiles      p1 ON p1.id = n.sender_id
JOIN   profiles      p2 ON p2.id = r.receiver_id
WHERE (r.receiver_id =  #{SELF.id} AND n.sender_id <> #{SELF.id} OR
       r.receiver_id <> #{SELF.id} AND n.sender_id =  #{SELF.id})
ORDER  BY c.id, c.updated_at DESC;

这会让您获得与每个群组中最新conversations.updated_at共享相同conversations.id的行。说明：
Select first row in each GROUP BY group?

第I列我不想分组=＆gt; COLUMN必须出现在GROUP BY子句中，或者在聚合FUNCTION中使用

1 个答案: