Obviously, indefinite integrals can differ by an arbitrary constant. But I haven't seen this before, and it looks neat. In fact one can integrate a/x for any constant a and come out with a = 0. Did you do this on your own or find it'somewhere else? I'm curious...