Rough sets are promising tools for expressing granular computer technology, with new investigation to come. Its representation of ability constraint, human thinking in the face of uncertainty, is incomprehensible to classical logic. This article studies eleven varieties of data using rough set theory. Using indiscernibility relation found that, classification definite by the seven sets of attributes is equivalence relation is same as classification definite by four sets attributes. Lower and upper, approximations and boundary region show result of varieties