using an aggregate value along with non-agrregate values in SQL - mysql

Question

I need to use the average value of the column 'sales' from a table called 'previous_target' and compare that value with the individual rows of the same column 'sales' of the same table .
I get the required result when I disable the SQL mode from only_full_group_by.
But i would like to know if there is a better way to write the code without disabling the full group by mode.
Here is an example SQL query.
select
f.id,t.sale as previous_target,
case
when t.sale > 1.25*(avg(t.sale)) then round((avg(t.sale)*1.1),2)
when t.sale < 0.9*(avg(t.sale)) then round((avg(t.sale)*0.9),2)
else
t.sale
end
as current_target from
details f
inner join prev_target t on f.l_number=t.l_number
inner join time_details ft on ft.id=f.id
note:if i add the line
group by f.id,f.l_number,t.sale
it just copies the same value onto the current_target column .
can anyone suggest a way to use the average of the sales column from the prev_target table and compare it with each row of the same table with the given conditions.
I hope I conveyed my requirement without causing much confusion.

Answer 1

SELECT f.id, t.sale AS previous_target,
CASE
WHEN t.sale > 1.25*(a.sale) then round(((a.sale)*1.1),2)
WHEN t.sale < 0.9*(a.sale) then round(((a.sale)*0.9),2)
ELSE
t.sale
END AS current_target
FROM details f
INNER JOIN prev_target t ON f.l_number = t.l_number
INNER JOIN
(
SELECT avg(t.sale) AS sale, t.l_number FROM prev_target t GROUP BY t.l_number
)AS a ON t.l_number = a.l_number
INNER JOIN time_details ft ON ft.id = f.id

Answer 2

Using OVER clause would be best recommended here for aggregation. I have created sample run for you (ofcourse different data and columns) but you will get the gist.
create table sales(amount int);
insert into sales values(5);
insert into sales values(10);
insert into sales values(15);
insert into sales values(20);
insert into sales values(10);
select * from sales;
select avg(amount) average from sales;
select amount, case when amount >= (avg(amount) over (PARTITION BY 1)) then 'Good' else 'Bad' end as type from sales;
Result >>