Why when using MIN function and selecting another column, we require GROUP BY clause? Doesn't MIN return single record? - sql

Question

I have two tables vehicles and dealership_vehicles. The dealership_vehicles table has a price column. The vehicles table has a column dealership_vehicle_id which relates to the dealership_vehicles id column in the dealership_vehicles table.
I wanted to return just the vehicle make of the cheapest car.
Why is it that the following query:
select
vehicles.make,
MIN(dealership_vehicles.price)
from
vehicles inner join dealership_vehicles
on vehicles.dealership_vehicle_id=dealership_vehicles.id;
Returns the error:
column "vehicles.make" must appear in the GROUP BY clause or be used in an aggregate function
Since MIN function returns a single value it is plausible that SQL query can be constructed that will return a single value without needing GROUP BY.

Answer 1

You say you want to know the make of the cheapest car. The easiest way to do this is
SELECT DISTINCT v.MAKE
FROM VEHICLE v
INNER JOIN DEALERSHIP_VEHICLES dv
ON v.DEALERSHIP_VEHICLE_ID = dv.ID
WHERE dv.PRICE = (SELECT MIN(PRICE) FROM DEALERSHIP_VEHICLES);
Note that because multiple vehicles might have the "cheapest" price it's entirely possible you'll get multiple returns from the above query.
Best of luck.
EDIT
Another way to do it is to take the minimum price, by make, then sort by the minimum price, and then just take the first row. Something like
SELECT *
FROM (SELECT v.MAKE, MIN(dv.PRICE)
FROM VEHICLE v
INNER JOIN DEALERSHIP_VEHICLES dv
ON v.DEALERSHIP_VEHICLE_ID = dv.ID
GROUP BY v.MAKE
ORDER BY MIN(dv.PRICE) ASC)
WHERE ROWNUM = 1;

Answer 2

Think of the term "GROUP BY" as "for each." It's saying "Give me the MIN of dealership_vehicles.price for each vehicles.make"
So you will need to change your query to:
select
vehicles.make,
MIN(dealership_vehicles.price)
from
vehicles inner join dealership_vehicles
on vehicles.dealership_vehicle_id=dealership_vehicles.id
Group by vehicles.make;

Answer 3

If you want the make of the cheapest car, then no aggregation is needed:
select v.make, dv.price
from vehicles v inner join
dealership_vehicles dv
on v.dealership_vehicle_id = dv.id
order by dv.price asc
fetch first one row only;
This gets a little more complicated if you want all rows in the case of ties:
select v.*
from (select v.make, dv.price, rank() over (order by price asc) as seqnum
from vehicles v inner join
dealership_vehicles dv
on v.dealership_vehicle_id = dv.id
) v
where seqnum = 1

Answer 4

So let's say after we join in the price we have the following table (i.e. stored in a #temp table):
#temp Vehicles table:
| Make | Model | Price |
|--------|-------|----------|
| Toyota | Yaris | 5000.00 |
| Toyota | Camry | 10000.00 |
| Ford | Focus | 7500.00 |
If you query it for minimum price without specifying what you're grouping by, then only one minimum function is applied across all of the rows. Example:
select min(Price) from #temp
will return you a single value of 5000.00
If you want to know the make of the cheapest car, you need to filter your results by the cheapest price - it's a two step process. First you find out the cheapest price using min, then in a separate query, you find out which cars are at that price. Once you construct your query correctly, you will notice that this reveals what you might not have though of - you can actually have more than one cheapest make.
Example table:
#temp Vehicles table v2:
| Make | Model | Price |
|--------|--------|----------|
| Toyota | Yaris | 5000.00 |
| Toyota | Camry | 10000.00 |
| Ford | Focus | 7500.00 |
| Ford | Escort | 5000.00 |
query:
select * from #temp
where Price = (select min(Price) from #temp)
result:
| Make | Model | Price |
|--------|--------|----------|
| Toyota | Yaris | 5000.00 |
| Ford | Escort | 5000.00 |