This paper provides a comprehensive survey of hybrid beamforming techniques in massive MIMO systems, which combine large-dimensional analog pre/postprocessing with lower-dimensional digital processing to reduce hardware cost and training overhead. The survey categorizes various hybrid beamforming structures based on the required channel state information (CSI), complexity, and carrier frequency range (cm-wave versus mm-wave). It discusses the challenges and solutions for each category, including the impact of CSI acquisition overhead, hardware imperfections, and the tradeoff between energy efficiency and spectral efficiency at mm-wave frequencies. The paper also highlights the importance of adapting these techniques to specific application and channel characteristics.This paper provides a comprehensive survey of hybrid beamforming techniques in massive MIMO systems, which combine large-dimensional analog pre/postprocessing with lower-dimensional digital processing to reduce hardware cost and training overhead. The survey categorizes various hybrid beamforming structures based on the required channel state information (CSI), complexity, and carrier frequency range (cm-wave versus mm-wave). It discusses the challenges and solutions for each category, including the impact of CSI acquisition overhead, hardware imperfections, and the tradeoff between energy efficiency and spectral efficiency at mm-wave frequencies. The paper also highlights the importance of adapting these techniques to specific application and channel characteristics.