Bypass mutating trigger in Oracle - sql

Question

I have two tables; on update of table 1 (cars bought), I need to update the other table with the sum of total cars bought by a specific customer.
When I try it, I encountered a mutating trigger. I have tried to convert it to a compound trigger, however I am encountered a vast number of errors, null index etc.
detail_table:
+-----------+---------+--------+------+------+
| customore | car | number | cost | sold |
+-----------+---------+--------+------+------+
| josh | mustang | 2 | 5 | y |
| josh | ford | 3 | 2 | y |
| josh | tesla | 1 | 3 | n | -->to update to y
| john | chevy | 4 | 1 | y |
| john | chevy | 5 | 2 | y |
+-----------+---------+--------+------+------+
On update of sold from n to y, the rows must roll up and sum into this summary table
summary_table
+----------+------------+------------+
| customer | total cars | total cost |
+----------+------------+------------+
| josh | 5 | 7 | -- > before update on detail
+----------+------------+------------+
+----------+------------+------------+
| customer | total cars | total cost |
+----------+------------+------------+
| josh | 6 | 10 | -- > after update on detail
+----------+------------+------------+
In the end when the user updates n to y for josh.total cars supposed to become 1 and total cost 10
trigger code
CREATE OR REPLACE TRIGGER update_summary_table
AFTER UPDATE ON detail_table FOR EACH ROW
referencing old as old new as new
begin
for i in (select sum(number) s_car, sum(cost) s_cost
from detail_table
where customer = :new.customer
and sold = 'y') loop
update summary_table
set total_cars = i.s_car,
total_cost = i.s_cost
where customer = :new.customer;
end loop;
end;
end update_summary_table;

Answer 1

Why would you use a for loop for this? Just use a single update:
update summary_table
set total_cars = coalesce(total_cars, 0) + :new.number - :old.number,
total_cost = coalesce(total_cost, 0) + :new.cost - :old.cost
where customer = :new.customer;
This does an incremental change to the summary rather than entirely recalculating the row.
The use of coalesce() is just in case no default value is given for new rows.