The underlying premise of your question is not true in general.
In the first place, you can never find the exact measurement of anything. ALL measurements are approximate, no matter how precisely taken. Therefore you can never exactly represent the surface area of a prism because the measurements of the edges begin as approximations.
On the other hand, who says you cannot represent the surface area of a cylinder exactly? Typically the surface area calculation for a cylinder involves , the ratio of the circumference of a circle to its diameter. is a transcendental irrational number, and irrational numbers can never be represented by an exact decimal fraction. However, that is not to say that you cannot represent a quantity involving exactly, because you can. All you need to do is leave the result in terms of and you have your exact representation. So presuming an ideal theoretical cylinder whose height was exactly and base radius exactly , you would have a lateral surface area of exactly or a total surface area of exactly . Example: If you had a cylinder that was purported to be exactly 5 units in height and exactly 2 units in radius (even though no one would ever be able to measure it and determine whether it was indeed 5 high and 2 in radius), then the lateral surface area would be exactly , and the total surface area would be exactly