You can choose to use the data type
FLOAT(p). When it is being used, the optional p specifies the minimum acceptable precision in binary digits and it can range between 1 to 53. FLOAT(1) to FLOAT(24) as selecting the REAL type, while FLOAT(25) to FLOAT(53) select DOUBLE PRECISION. A value of p outside the allowed range results in an error. The default value is 53 (maximum precision).