This paper studies the Graph-Connected Clique-Partitioning Problem (GCCP), a clustering optimization model in which units are characterized by both individual and relational data. This problem, introduced by Benati et al. ( 2017) under the name of Connected Partitioning Problem, shows that the combination of the two data types improves the clustering quality in comparison with other methodologies. Nevertheless, the resulting optimization problem is difficult to solve; only small-sized instances can be solved exactly, large-sized instances require the application of heuristic algorithms. In this paper we improve the exact and the heuristic algorithms previously proposed. Here, we provide a new Integer Linear Programming (ILP) formulation, that solves larger instances, but at the cost of using an exponential number of variables. In order to limit the number of variables necessary to calculate the optimum, the new ILP formulation is solved implementing a branch-and-price (B&P) algorithm. The resulting pricing problem is itself a new combinatorial model: the Maximum-weighted Graph-Connected Single-Clique problem (MGCSC), that we solve testing various Mixed Integer Linear Programming (MILP) formulations and proposing a new fast "random shrink" heuristic. In this way, we are able to improve the previous algorithms: The B&P method outperforms the computational times of the previous MILP algorithms and the new random shrink heuristic, when applied to GCCP, is both faster and more accurate than the previous heuristic methods.Moreover, the combination of column generation and random shrink is itself a new MILPrelaxed matheuristic that can be applied to large instances too. Its main advantage is that all heuristic local optima are combined together in a restricted MILP, consisting in the application of the exact B&P method but solving heuristically the pricing problem.