sql查询组按特定结果作为两个字段值的组合值

Question

My message table contains two fields, senderid and recipientid . I need to list all messages but the combination of values in senderid and recipientid don't need to repeat.

For example if senderid = 4 and recipientid = 2 , I need only one combination value of this; either it is 4,2 or its is 2,4.

Can any one help me in doing this?

Thanks in advance ;)

Answer 1

select
    message,
    case when senderid < recipientid then recipientid else senderid end,
    case when senderid < recipientid then senderid else recipientid end
from m
group by
    message,
    case when senderid < recipientid then recipientid else senderid end,
    case when senderid < recipientid then senderid else recipientid end

Answer 2

select *
from myTable a
where not exists
(
    select top 1 1 
    from myTable b
    where a.senderId = b.receiverId
    and a.receiverId = b.senderId
    and a.senderId < b.senderId
)

Answer 3

Idea is simple, just make sure the sender/recipient IDs are concatenated in the way such that the smaller IDs is in front of the larger IDs.

select
  pair, count(*) 
from 
  (select if(senderid<recipientid, concat(senderid, ',', recipientid), concat(recipientid,',',senderid)) as pair from messages) pairs
group by pair;

Mind the performance of such selection. Maybe you would like to create an additional column to store the pair value when inserting the record, and have the column indexed.