正如tidoss所提到的。某种窗口函数。这里有一个带行号的函数，假设你的id可以用来确定什么是“最后一个”

with t as
 (select m.*,
  row_number() over (partition by userid order by id desc) as rn
  from mytable m
)
select id, userid, score from t where rn = 1

解决此问题的类似方法是，首先按userid的升序和id的降序对表进行排序，然后按外部userid对inner选择的值进行分组
内部选择

SELECT inner.* FROM table AS inner ORDER BY userid ASC, id DESC;

会给予这个中间结果：

id  userid  score
--  ------  -----
3   10      70   
2   10      89   
1   10      55   
5   15      80   
4   15      50   
7   17      99   
6   17      90

按userid对该结果进行分组将给予您想要的结果。

SELECT outer.* FROM 
  (SELECT inner.* FROM table AS inner ORDER BY userid ASC, id DESC) AS outer
GROUP BY outer.userid;

这将为您提供：

id  userid  score
--  ------  -----
3   10      70   
5   15      80   
7   17      99