The retail company sales vegetables.

Total company's costs are divided on 10 groups (positions) such as
employee's compensation,
transport, taxes etc. The costs are also divided between all
vegetables (each good has different cost
for each cost position)

I have to tables in my universe:
1) Sales: veg_name, revenue
2) Costs: veg_name, cost_position, cost_value
And some data:
1)
Tomatoes 10000
Potatoes 20000
Carrots 9000
2)
Tomatoes Taxes 100
Tomatoes Transport 200
Tomatoes Emp. compensation 130
Carrots Taxes 100
Carrots Transport 200
....
and so on

I've designed the universe with 2 classes: SALES and COSTS
Sales includes VEG_NAME as dimension and REVENUE as value (sum for
aggregation).
Costs includes VEG_NAME as dimension, COST_POSITION as dimension and
COST_VALUE as value
(sum for aggregation).

Ok, I've created the report containing all variables from universe:
VEG_NAME, REVENUE, COST_POSITION, COST_VALUE

Now I have a problem: if I'll cut the COST_POSITION dimension then
I'll get
double counting for REVENUE.
How should I reorganize my universe to get the sum(REVENUE) invariable
while cutting
the COST_POSITION ?