It is non-trivial to use group cross-section data, that has been taken from a cross-section library, directly in a computer code. In this document two different methods to process data stored as equi probable cosines bins have been developed. The first method determines the Legendre moments of the scattering cross-section in a simple way. In some cases, however, it may lead to a negative cross-section in some directions. This usually happens when the scattering is highly anisotropic. One consequence of that is that the neutron flux may be negative, which is unacceptable. The second method is based on a smoothing technique using a window function. In general this method solves the problems that is inherent in the first method. There is a potential problem with this method caused by the truncation of a Legendre expansion of the window function. However, for many practical cases the errors made are expected to be small.