SQL: Efficient way to get group by results including all table columns

Question

Let's consider a simple table below.

id	code	marks	grade
1	X	100	A
2	Y	120	B
3	Z	130	A
4	X	120	C
5	Y	100	A
6	Z	110	B
7	X	150	A
8	X	140	C

Goal: Get maximum marks for each grade, return all the columns.

id	code	marks	grade
7	X	150	A
2	Y	120	B
8	X	140	C

This is very simple if I don't want id and code column

select grade, max(marks)
from table
group by grade;

What could be the most efficient query to get id and code column in the above query?

I tried something like this which didn't work

select * from table t
inner join
(select grade, max(marks)
from table
group by grade) a
on a.grade=t.grade;

Answer 1

In Postgres the most efficient way for this kind of query is to use (the proprietary) distinct on ()

select distinct on (grade) *
from the_table t
order by grade, marks desc;