SQL - How to retrieve the rows with the highest rank on a column with multiple duplicates rows in another one - sql

Question

I'm quite new to SQL, so it might be a noob question:
Let's say my query is:
select [Item], [Answers] from table1
And it retrieves the following table:
Item Answers
------------------
Car Expensive
Car Cheap
Car Medium
Boat Expensive
Boat Very Expensive
Boat Ultra Expensive
Given a certain second table (or dictionary, I don't really know how to do it) {Cheap: 1, Medium: 2, Expensive: 3, Very Expensive:4, Ultra Expensive: 5} - meaning that "Ultra Expensive" is the highest rank and "Cheap" is the lowest rank.
In SQL, on this kind of table with many duplicates in Column A (Item) how do I retrieve the highest ranked value in Column B (Answers) for each unique value in Column A?
In this example, I would like to get:
Item Answers
------------------
Car Expensive
Boat Ultra Expensive
Just one of each duplicated value in column 'Item' and its highest ranked possible value in 'Answers'?

Answer 1

You can use a correlated subquery:
select item, answer
from table1 t
where answer = (
select top 1 answer
from table1 t1
where t1.item = t.item
order by
case answer
when 'Cheap' then 1
when 'Medium' then 2
when 'Expensive' then 3
when 'Very Expensive' then 4
when 'Ultra Expensive' then 5
end desc
)
The subquery filters on the answer that has the highest rank for the given item, using a conditional order by clause and top 1.
Another option is to use row_number() for filtering:
select item, answer
from (
select
item,
answer,
row_number() over(
partition by item
order by case answer
when 'Cheap' then 1
when 'Medium' then 2
when 'Expensive' then 3
when 'Very Expensive' then 4
when 'Ultra Expensive' then 5
end desc
) rn
from table1
) t
where rn = 1