Computer systems use internal registers to performs math functions.
These registers are typically 32-bit, 64-bit, 128-bit, or 256-bit in length.
DOUBLE PRECISION gives a double-precision floating-point number (the size of two registers).

> The actual precision of DOUBLE PRECISION depends on the RDBMS vendor’s implementation.

> Since it is a floating-point number expressed in scientific notation, it is an approximate number and may not be accurate to the last digit.

> It can be negative, zero, or positive.

> If you have a choice between exact (DECIMAL or NUMERIC) or approximate (FLOAT, REAL, or DOUBLE PRECISION), use the exact data type because you get completely accurate numbers and it processes faster.

> SQL:2003 introduced REAL, DOUBLE PRECISION, and FLOAT(precisionbits) data types.
> Approximate data type should be used only for very extreme ranges of numeric values.

Computer systems use internal registers to performs math functions.
These registers are typically 32-bit, 64-bit, 128-bit, or 256-bit in length.
REAL gives a single-precision floating-point number (the size of one register).

> The actual precision of REAL depends on the RDBMS vendor’s implementation.

> Since it is a floating-point number expressed in scientific notation, it is an approximate number and may not be accurate to the last digit.

> It can be negative, zero, or positive.

> If you have a choice between exact (DECIMAL or NUMERIC) or approximate (FLOAT, REAL, or DOUBLE PRECISION), use the exact data type because you get completely accurate numbers and it processes faster.

> SQL:2003 introduced REAL, DOUBLE PRECISION, and FLOAT(precisionbits) data types.
> Approximate data type should be used only for very extreme ranges of numeric values.

FLOAT is a floating point number expressed in scientific notation. It is a decimal number multiplied by an integer power of 10. For example: 4.5E2 = 4.5 x 10 = 450. The mantissa (coefficient) is the portion that expresses the significant digits (here it’s 4.5). An upper case E is the symbol for the exponent (here it’s 2 for 10 to the 2nd power). The mantissa and exponent can also be negative. For example: -4.5E-2 = -4.5 x 10 = -0.045.

Example:    COL1 FLOAT(24)

COL1 can have the precision of 24 bits or ten million significant digits.

Definition  Possible Values                    Example
FLOAT(12)    12 bits or 3 significant digits   3.14E2 = 314
FLOAT(24)    24 bits or 7 significant digits   5.280796E0 = 5.280796
FLOAT(34)    34 bits or 10 significant digits  3.297249382E-3 = 0.003297249382
FLOAT(50)    50 bits or 15 significant digits  1.45732792216439E11
FLOAT(100)  100 bits or 30 significant digits

> Multiply the bits precision by 0.30103 to get the decimal precision.
> Multiply the decimal precision by 3.32193 to get the binary precision.
> For most RDBMS systems the range is -5.4E-79 to +7.2E+75.
> If precisionbits is 1 to 21, then the column value is 5 bytes long.
    If precisionbits is 22 to 53, then the column value is 9 bytes long.