Postgresql select top 5 marks rows for every unique student id - database

Question

This is one of my views called xyz:
ID | NAME |......Other Data... | Marks
1 A 100
1 A 100
1 A 99
1 A 95
1 A 94
1 A 94
1 A 94
1 A 91
1 A 87
1 A 86
2 B 100
2 B 94
2 B 93
2 B 90
2 B 89
2 B 89
2 B 87
2 B 86
3 C 100
3 C 98
3 C 98
3 C 97
3 C 92
3 C 91
3 C 90
The query I used to get this is something like this:
create or replace view xyz as
select *
from abc
where id in
(select id
from data)
order by id, mark desc, id;
And based on unique ids I want the top 5 rows:
ID | NAME |......Other Data... | Marks
1 A 100
1 A 100
1 A 99
1 A 95
1 A 94
2 B 100
2 B 94
2 B 93
2 B 90
2 B 89
3 C 100
3 C 98
3 C 98
3 C 97
3 C 92
I tried referring to this:
MySQL: Select top 5 rows based on ID and find Subtotal
But I wasn't able to do it. Could you please help?

Answer 1

You can use a windowed rank(), partitioned by each student name, and ordered by marks descending, to find the top 5 marks for each Student:
WITH cteRankedMarks AS
(
SELECT "ID", "NAME", "Marks",
rank() OVER (PARTITION BY "NAME"
ORDER BY "Marks" DESC) AS rank
FROM MyTable
)
SELECT "ID", "NAME", "Marks"
FROM cteRankedMarks
WHERE rank <= 5
ORDER BY "NAME", "Marks" DESC;
SqlFiddle here
Notes
If two or more marks have the same value vying for 5th place, then , then rank will return all such marks. If you don't want ties, then use row_number() instead of rank(). More about these here
Obviously, if a student doesn't have 5 marks, then fewer rows will be returned.