Of course, a very very eccentric ellipse approximates a parabola quite well. You only start to notice errors when dealing with trajectories dozens of miles long (on Earth).