Six problem formulations exist in the multi-criteria decision analysis (MCDA): choice, sorting, ranking, description, elimination and design problems. The analytic hierarchy process (AHP) is a useful and widespread method for solving choice and ranking problems. However, it is not adapted for sorting problems. Moreover, another practical limitation of AHP is that a high number of alternatives implies a large number of comparisons. This paper presents AHPSort, a new variant of AHP, used for the sorting of alternatives into predefined ordered categories. Furthermore, AHPSort requires far less comparisons than AHP, which facilitates decision making within large-scale problems. In this paper, a real case study for supplier selection is used to illustrate our approach. First, the candidates are sorted with AHPSort within two classes: accepted and rejected suppliers. Then, a single supplier is selected with AHP among the accepted suppliers.