Is there any difference between 1D or 2D array input to a simple neural network?

Question

I am training a DQN with previous two observations as input to the DQN. My model consists of only one Dense layer with 20 units and the output is the probabilities