There are four floating-point types that are accepted by the preprocessor. The types float and f_float are the 4-byte floating-point types. The types long_float and d_float are the 8-byte floating-point types. Long_float requires some extra definitions or default Ada pragmas to be able to interact with Ingres floating-point types. Note that the preprocessor does not accept long_long_float and h_float data types.