Floating-Point Types
AttentionThis page documents an earlier version. Go to the latest (v2.1)version.
Synopsis
Floating-point and fixed-point numbers are used to specify non-integer numbers. Different floating point datatypes represent different precisions numbers.
DataType | Description | Decimal Precision |
---|---|---|
REAL | Inexact 32-bit floating point number | 6 |
DOUBLE PRECISION | Inexact 64-bit floating point number | 15 |
FLOAT | Inexact 64-bit floating point number | variable |
DECIMAL | Exact fixed-point number | 99 |
Syntax
type_specification ::= { FLOAT | DOUBLE PRECISION | DECIMAL | DEC }
non_integer_floating_point_literal ::= non_integer_fixed_point_literal | "NaN" | "Infinity" | "-Infinity"
non_integer_fixed_point_literal ::= [ + | - ] { digit [ digit ...] '.' [ digit ...] | '.' digit [ digit ...] }
Where
- Columns of type
FLOAT
,DOUBLE PRECISION
,DEC
, orDECIMAL
can be part of thePRIMARY KEY
. non_integer_floating_point_literal
is used for values ofFLOAT
,DOUBLE
andDOUBLE PRECISION
types.non_integer_fixed_point_literal
is used for values ofDECIMAL
type.
Semantics
- Values of different floating-point and fixed-point datatypes are comparable and convertible to one another.
- Conversion from floating-point types into
DECIMAL
will raise an error for the special valuesNaN
,Infinity
, and-Infinity
. - The ordering for special floating-point values is defined as (in ascending order):
-Infinity
, all negative values in order, all positive values in order,Infinity
, andNaN
. - Values of non-integer numeric datatypes are neither comparable nor convertible to integer although integers are convertible to them. This restriction will be removed.
See Also
当前内容版权归 YugabyteDB 或其关联方所有,如需对内容或内容相关联开源项目进行关注与资助,请访问 YugabyteDB .