This letter investigates the coexistence between near-field (NF) and far-field (FF) communications, where multiple FF users are clustered to be served on the beams of legacy NF users, via non-orthogonal multiple access (NOMA). Three different successive interference cancellation (SIC) decoding strategies are proposed and a sum rate maximization problem is formulated to optimize the assignment and decoding order. The beam allocation problem is further reformulated as an overlapping coalitional game, which facilitates the the design of the proposed clustering algorithm. The optimal decoding order in each cluster is also derived, which can be integrated into the proposed clustering. Simulation results demonstrate that the proposed clustering algorithm is able to significantly improve the sum rate of the considered system, and the developed strategies achieve different trade-offs between sum rate and fairness.