FLOAT

The 32-bit single-precision floating-point type (FLOAT) follows the IEEE 754 standard, with an effective precision of approximately 6 to 7 decimal digits. It is suitable for storing real numbers that do not require exact computation. For exact computation scenarios, use DECIMAL.

Syntax

FLOAT REAL

REAL is an alias for FLOAT, provided for compatibility with migration scripts from other databases. Aliases are immediately converted to the canonical type during parsing; see Type Aliases for details.

Value Range

• Maximum positive value: approximately 3.4028235 × 10³⁸
• Minimum positive value (non-zero): approximately 1.4 × 10⁻⁴⁵
• Literal suffix: F (e.g., 1.5F, -3.2F)

Examples

1. Using the FLOAT literal suffix:

   SELECT 1.5F;

   Returns: 1.5

2. Convert an integer to FLOAT:

   SELECT CAST(6 AS FLOAT);

   Returns: 6.0

3. Convert a string to FLOAT:

   SELECT CAST('3.14' AS FLOAT);

   Returns: 3.14

4. Precision loss example:

   SELECT CAST(1234567.89 AS FLOAT);

   Returns: 1234568.0 (exceeds 7 significant digits, precision loss occurs)

5. NULL value handling:

   SELECT CAST(NULL AS FLOAT);

   Returns: NULL

Type Selection Guide

FLOAT vs DOUBLE precision comparison (measured):

SELECT CAST(3.14159265358979 AS FLOAT) AS f, -- 3.1415927410125732 (distortion after 7 digits) CAST(3.14159265358979 AS DOUBLE) AS d; -- 3.14159265358979 (fully preserved)

Notes

• FLOAT is an approximate numeric type with effective precision of about 6-7 digits. It is not suitable for scenarios requiring exact computation such as financial amounts; use DECIMAL instead.
• Avoid using = to directly compare two FLOAT values; precision errors may produce unexpected results.
• Arithmetic overflow returns Infinity or -Infinity; invalid operations such as sqrt(-1) return NaN.
• CAST of an invalid string returns NULL.