Question

我有一个名为messages的SQL表，其中包含以下详细信息;

id, sender, recipient, message, timestamp

"1","10204456572654160","10208819391147662","Hi.. : What are you up to?","1459541723279"
"2","10204456572654160","10208819391147662","Got my test message?","1459541749818"
"3","10208819391147662","10204456572654160","This is my message","1459611679108"
"4","10208819391147662","10204456572654160","And another message","1459611735455"
"5","10204456572654160","10208819391147662","And I reply like this","1459611758570"
"6","10153775515332771","10208819391147662","It's me.. Syco !!","1459611900348"
"7","10153775515332771","10208819391147662","You there?","1459611900350"
"8","10208819391147662","10153775515332771","Yes.. What's upp..","1459611900380"
"9","10204464403169266","10208819391147662","How're you doin?","1459612000666"

现在，我想将一个值传递给我的函数e.g. 10208819391147662并显示结果，如下面的列;

sender/recipient的唯一值不 10208819391147662
message之间的最后sender/recipient 不 10208819391147662 AND sender/recipient 10208819391147662 }。
第二个timestamp（消息）。

为了实现以下目标，我有这个sql;

SELECT s1.*, max(c2.message), max(c2.timestamp)
FROM (
    SELECT sender as username 
    FROM messages
    WHERE sender <> '10208819391147662'
    UNION
    SELECT recipient as username 
    FROM messages
    WHERE recipient <> '10208819391147662'
) s1
LEFT JOIN messages c2 ON (s1.username = c2.sender OR s1.username = c2.recipient)
GROUP BY s1.username

我已成功使用1st column提取UNION。但是使用MAX和Group by子句在第2和第3位没有运气。

我的最终预期结果应该是;

username, message, timestamp

"10204456572654160","And I reply like this","1459611758570"
"10153775515332771","Yes.. What's upp..","1459611900380"
"10204464403169266","How're you doin?","1459612000666"

到目前为止我的SQLFiddle了。任何建议都表示赞赏。

Answer 1

实际上，最后这是我的最终SQL，它完成了我想要的东西。

SELECT c2.id, s1.username, c2.message, max(c2.timestamp)
FROM (
    SELECT sender as username 
    FROM messages
    WHERE sender <> '10208819391147662'
    UNION
    SELECT recipient as username 
    FROM messages
    WHERE recipient <> '10208819391147662'
) s1
LEFT JOIN messages c2 ON (s1.username = c2.sender OR s1.username = c2.recipient)
GROUP BY s1.username
ORDER BY timestamp DESC

但不知何故，在SQLFiddle中，这在MYSQL中无法正常工作，但在SQLITE中效果很好。

SQLITE查询连接两个具有不同条件的相同表

1 个答案: