-1

I have a query from a mysql table that results in a dataset as follows

car colour sold
Benz Black 2
Benz White 1
BMW Black 3
BMW Green 2
BMW Blue 1
Toyota Yellow 7
Toyota Red 5
Toyota Blue 3
Toyota White 1

I'm trying to select the top sold row from each subrow, so my final dataset would be:

car colour sold
Benz Black 2
BMW Black 3
Toyota Yellow 7

Order doesn't matter i just need to make sure its the top colour sold for that particular car. I tried using distinct, but that applies to the whole row. I tried using group by that selects a random sold amount not the top one. Any idea what query I should be running for this? Is using subqueries inevitable? its a rather small dataset of about ~100 entries. But id rather not use subqueries for future scaling.

Oirampok
  • 569
  • 5
  • 23

1 Answers1

0

We can use the function RANK()in a CTE.
See the dbFiddle link at the bottom for the schema and further testing.
NB: This will not function with older version of mySQL.

with rankedSales as (
select 
  car,
  colour,
  sold,
  rank() over (partition by car order by sold desc) rn
from carSales)
select
  car, 
  colour,
  sold
from rankedSales
where rn = 1;
car    | colour | sold
:----- | :----- | ---:
Benz   | Black  |    2
BMW    | Black  |    3
Toyota | Yellow |    7

db<>fiddle here