On top of what Lonnie Said;
Sometimes flow doesnt just appear at the channel of the stream and if you enter anything like 0.01 the flow will just drop into the channel (Cell with the lowest elevation) without wetting any other cells. one way around this is to distribute that line across the area of inflow i.e. the top of a mountain range, and apply an extremely LOW slope like 1e-5 and lower to 1e-10 (yes HEC-RAS reads inputs of 1e-10 without the need to hit zero a bunch of times)
A colleague and i went through extreme amounts of testing putting different energy slopes until we said lets try something VERY low and lets try something VERY high. and voila, we saw what the software was doing.
What you will notice is that each cell now becomes wet upon entry into the hydraulic model and you can watch as the flow actually channelizes. This Will make your model think harder but in many cases this is better than measuring the slope along the face of your 2D area and inputting that slope; because EVERYTIME you do this, the flow will just populate at the lowest cell elevation along that line and not distribute across the cells for which you have drawn them.