Return the indices of the non-zero elements of the 1-D array a.
a
solve([1, 2, 0, 0, 4, 0]) -> array([0, 1, 4])
solve(...)