Published 24 November 2017
FLOAT/REAL datatype is used
The FLOAT (8 byte) and REAL (4 byte) data types are suitable only for specialist scientific use since they are approximate types with an enormous range (-1.79E+308 to -2.23E-308, and 2.23E-308 to 1.79E+308, in the case of FLOAT).
Any other use needs to be regarded as suspect; a FLOAT or REAL used as a key or found in an index needs to be investigated.
The DECIMAL type is a fixed-precision data type with a range from -10^38+1 through 10^38-1. Although it requires more storage than the FLOAT or REAL types, it is generally a better choice.