Toby et. al (2003) recommend looping phase information in the "diffraction data block", and include _pd_proc_ls.profile_function in the tabulation, that is, the profile function could be phase-dependent. They note that the data name definition will therefore need to be fixed in the future.
With preferred orientation we have deprecated and created a whole new category, which would seem to be the right approach here too.