I have two arrays, and I am trying to find the displacement for each elements to make them match. An example: a = np.array([1, 0, 2, 0, 0, 1, 0, 2, 0, 0, 1, 0, 2, 0, 0 ] ) b = np.array([2, 0, 0, 1, 0, 2, 0, 0, 1, 0, 2, 0, 0, 1, 0] ) # expected c = np.array ([3, 0, 2, 1, 0, 2, 0, 2, 1, 0, 2, 0,2, 1, 0 ]) The expected results, in array a, the first element is 1, and the closest match in b is 3 steps away. The second element in a is 0 and it is a match in b so zero steps away. The third element is 2, and the closes in b is b[0] which is 2 steps away from a[2]=2 .. and so on. I tried to come up with a function to solve this: def find_shifts_required(x, y): assert len(x) == len(y) result = np.zeros(len(x)) labels = np.unique(x) for lbl in labels: x_pos = np.where(x == lbl) y_pos = np.where(y == lbl) #.min()) result[x_pos] = np.subtract(x_pos , y_pos ) _shifts = np.absolute(result).astype(int) return _shifts d = find_shifts_required(a,b) print("expected: " , c) print("actual: " , d) Which gives me: expected: [3 0 2 1 0 2 0 2 1 0 2 0 2 1 0] actual: [3 0 2 1 0 3 0 2 1 0 3 0 2 1 0] I am wondering if this problem has been solved somewhere (scipy: I wasn't able to find a name) or a better implementation is available. By better I mean a working solution, as this one does not give the correct results. A simpler, faster will be great. Continue reading...
