We show how fuzzy and rough data are transformed into a unified representation using the concept of rough fuzzy sets. We introduced a representation for rough fuzzy classifications that might be used when the classification process introduces uncertainty in the data due to vagueness and indiscernibility. We demonstrate the viability of our method by performing classification experiments using real data. The experiment uses data that have been classified using different classification schemes. To make them compatible we reclassify some data and use some additional knowledge that requires the use of rough fuzzy classifications.