But when your input is 5 kg * 1 EUR/kg, you have 5€ of costs to produce the table. And if the table has a revenue of 33 EUR in the output, you have an added value of 28€. All good, no?
If you would set all market values of the quantitative references to 0, you would not get any result. Because the next process in the supply chain will not know what the costs of the inputs are.
Via script it is possible to write all prices into the inputs and remove all the prices in the outputs, but I don't know what will be the use of it, since you would add up all thousands of inputs of a complex product system in ecoinvent, which will lead to a very high, unreasonalbe result. Just imagine a market process where a product flow "hikes" through. You will double count all these input prices in processes.
An advise that I can give you is maybe removing all prices from ecoinvent and only do the LCC for your own foreground and manually for the linked ecoinvent processes to your model. This will be more accurate anyway. So you build your own foreground model where you give prices to everything, but you do not add prices inside "proctected" ecoinvent processes.
You can remove all prices (do a backup database before) with SQL:
UPDATE TBL_EXCHANGES SET COST_VALUE = NULL;