We consider the opportunistic multiuser diversity in the multiuser two-way amplify-and-forward (AF) relay channel. The relay, which is equipped with multiple antennas and a simple zero-forcing beamforming (ZFBF) scheme, selects a set of two way relaying user pairs to enhance the degree of freedom (DoF) and, consequently, the sum throughput of the system. The proposed channel-aligned pair scheduling (CAPS) algorithm reduces the interpair interference and keeps the signal-to-interference-plus-noise-power ratio (SINR) of user pairs that are interference-free when the number of user pairs becomes very large. When the number of user pairs grows fast enough with the system signal-to-noise ratio (SNR), a DoF equal to the number of relay antennas can be achieved. For a realistic number of user pairs, we propose an adaptive CAPS and an adaptive semi-orthogonal CAPS (SCAPS) to improve performance. Simulation results show that adaptive CAPS and adaptive SCAPS provide throughput gain in the low-to-mid SNR region.