Spark data processing for deltas over time

Question

My dataset columns looks like device_name date categorycol1 categorycol2 .. categorycol10 valuecol1