Message
模型
user_id # owner of the message
sender_id # user that sent message
receiver_id # user then recieved message
content # content of message
有Messages
。我想把它们按sender_id
和receiver_id
分组,因为它们是一个"对话线程"。
当我执行group时,我得到的结果看起来像
[1,3] => 5 # user 1 and 3 have 5 messages
[1,6] => 2 # user 1 and 6 have 2 messages
[3,1] => 3 # user 3 and 1 have 3 messages
真的[1,3]和[3,1]是同一个"组"的一部分。我怎样才能做到这一点呢?
你可以通过使用LEAST()
和GREATEST()
来排序你的sender_id
和recipient_id
(至少在Postgres和MySQL中)。
下面是演示SQL:
SELECT
LEAST(sender_id, receiver_id),
GREATEST(sender_id, receiver_id),
count(*)
FROM messages
GROUP BY
LEAST(sender_id, receiver_id),
GREATEST(sender_id, receiver_id)
和对应的ActiveRecord
:
Message.group('LEAST(sender_id, receiver_id), GREATEST(sender_id, receiver_id)').count
使用@Kristján提供的技巧,您可以在ruby中合并[1,3]
和[1,3]
,而不是使用SQL函数:
thread_messages_count = Hash.new(0)
Message.group(:sender_id, :receiver_id).count.each do |k, v|
thread_messages_count[k.sort] += v
end