Question

I am trying to write a query for a large dataset with many joins and having trouble accomplishing a particular piece without some sort of subquery, which I am trying to avoid.

For an example table with columns ID, Size, Item there may be multiple records with the same ID. I want to return the record per ID which has the largest Size.

ID  Size    Item
1   5   a
1   10  b
2   3   c
2   6   d
2   11  e
3   2   f

Expected result

ID  Size    Item
1   10  b
2   11  e
3   2   f

I’ve tried various group and having approaches without success.

Using a subquery I can do it like this but for a large dataset I’d prefer not to do it this way

select id, size, item
from test
where size = (select max(size) from test t2 where id = test.id)

Any suggestions?

Answer 1

This should satisfy your requirements: For each id, return only the row with the largest size

SELECT test.id, test.size, test.item
FROM test
INNER JOIN (
    SELECT id, MAX(size) AS size
    FROM test
    GROUP BY id
) max_size ON max_size.id = test.id AND max_size.size = test.size