In this paper, we develop novel precoders and combiners for multi-user hybrid digital to analog (D-A) beamforming for uplink massive multiple-input multiple-output millimeter wave systems. Considering the possibility that uplink transmissions from different users can go through the paths sharing the same physical scatters, some transmission paths of different users may have overlapped angle of arrivals (AoAs) at the base station. Therefore, the intrinsic focus is on substantially maximizing the desired signal while reducing the system interference. The analog precoders and combiners are designed by using the eigenvalue decomposition and the Riemannian optimization method based on Stiefel manifold algorithms respectively. The digital combiner adapts the minimum mean-square error by judiciously exploiting the effective uplink analog channel gains. Ultimately, simulation results show that the proposed algorithms outperform the existing algorithms in hybrid D-A paradigms in terms of the achievable uplink rate.