Basic nuclear data libraries, such as ENDF/B-III, are used today for the generation of multigroup cross-section sets. Since the multigroup sets are determined by calculating the weighted averages of the cross-section curves, knowledge of the weighting function within each group is required. In principle, this function is not only dependent on the specific reactor, but is also a function of position in the reactor. With existing codes and methods the weighting function in the Bondarenko-type library is not iterated, since each time the weighting function is changed, a code that handles the basic data library and consumes much computer time must be run.