Three ways to write of group by xxx about each user taking 1 record in MYSQL

Today, some students asked me about the SQL statement in this aspect. I specially memorized it for 1, because it is also common

select * from (select * from member_payment 
                order by id desc) t group by member_id limit 10

The first way is to sort first, then group, so that you can get the most suitable data.
The disadvantages are obvious: Using temporary; Using filesort

select s.* 
from (SELECT max(id) as id FROM `member_payment` group by `member_id` limit 10) t 
left join `member_payment` as s on t.id=s.id 

The second is a joint query

select * from `member_payment` where EXISTS (
 select `id` from (
  SELECT max(`id`) as id FROM `member_payment` group by `member_id` limit 10) t 
 where t.`id`=`member_payment`.`id`
)

The third type is subqueries
I think the second one is the most efficient