we've got information twin opposite origins: comes customer, comes opposite vendors. currently, physically "merge" information large list roughly hundred columns, tens thousands rows grave fortitude twin dimensions. consequently, can't indeed list much. i'm going redesign disaster proper, nonetheless small, star schema. the twin magnitude obvious. them, example, time. the customer-supplied information provides array fact values. any businessman competence (or competence not) yield additional fact values fit same dimensions. this fact information same granularity. called "sparse" since don't mostly information vendors. here's dilemma. is fact list -- nulls -- populated opposite sources? or n +1 fact tables -- populated customer, others populated any vendor? there pros cons any design. i need second opinions choice between "merge" "load separately". customer fin revenue, cost, counts, weights, things know finish transa...