A floating-point value is represented either as whole plus fractional digits (for example, 123.45) or as a mantissa plus an exponent. The following figure illustrates the mantissa and exponent parts of floating point values.

There are two floating point data types: float4 (4-byte) and float (8-byte). (Real is a synonym for float4, and float8 and double precision are synonyms for float.)

Floating point numbers are double-precision quantities stored in four or eight bytes. The range of float values is processor-dependent, and the precision is approximately 16 significant digits.

For information about the correct notation for a floating-point numeric literal, see Numeric Literals.