Enum DataType

#[non_exhaustive]
pub enum DataType {
    Bool,
    Int2,
    Int4,
    Int8,
    Int16,
    Int32,
    Int64,
    UInt2,
    UInt4,
    UInt8,
    UInt16,
    UInt32,
    UInt64,
    Float4E2M1FN,
    Float6E2M3FN,
    Float6E3M2FN,
    Float8E3M4,
    Float8E4M3,
    Float8E4M3B11FNUZ,
    Float8E4M3FNUZ,
    Float8E5M2,
    Float8E5M2FNUZ,
    Float8E8M0FNU,
    BFloat16,
    Float16,
    Float32,
    Float64,
    ComplexBFloat16,
    ComplexFloat16,
    ComplexFloat32,
    ComplexFloat64,
    ComplexFloat4E2M1FN,
    ComplexFloat6E2M3FN,
    ComplexFloat6E3M2FN,
    ComplexFloat8E3M4,
    ComplexFloat8E4M3,
    ComplexFloat8E4M3B11FNUZ,
    ComplexFloat8E4M3FNUZ,
    ComplexFloat8E5M2,
    ComplexFloat8E5M2FNUZ,
    ComplexFloat8E8M0FNU,
    Complex64,
    Complex128,
    RawBits(usize),
    String,
    Bytes,
    NumpyDateTime64 {
        unit: NumpyTimeUnit,
        scale_factor: NonZeroU32,
    },
    NumpyTimeDelta64 {
        unit: NumpyTimeUnit,
        scale_factor: NonZeroU32,
    },
    Extension(Arc<dyn DataTypeExtension>),
}

A data type.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Bool

bool Boolean.

Int2

int2 Integer in [-2, 1].

Int4

int4 Integer in [-8, 7].

Int8

int8 Integer in [-2^7, 2^7-1].

Int16

int16 Integer in [-2^15, 2^15-1].

Int32

int32 Integer in [-2^31, 2^31-1].

Int64

int64 Integer in [-2^63, 2^63-1].

UInt2

uint2 Integer in [0, 3].

UInt4

uint4 Integer in [0, 15].

UInt8

uint8 Integer in [0, 2^8-1].

UInt16

uint16 Integer in [0, 2^16-1].

UInt32

uint32 Integer in [0, 2^32-1].

UInt64

uint64 Integer in [0, 2^64-1].

Float4E2M1FN

float4_e2m1fn a 4-bit floating point representation: sign bit, 2 bit exponent (bias 1), 1 bit mantissa.

• Extended range: no infinity, no NaN.
• Subnormal numbers when biased exponent is 0.

Float6E2M3FN

float6_e2m3fn a 6-bit floating point representation: sign bit, 2 bit exponent (bias 1), 3 bit mantissa.

• Extended range: no infinity, no NaN.
• Subnormal numbers when biased exponent is 0.

Float6E3M2FN

float6_e3m2fn a 6-bit floating point representation: sign bit, 3 bit exponent (bias 3), 2 bit mantissa.

• Extended range: no infinity, no NaN.
• Subnormal numbers when biased exponent is 0.

Float8E3M4

float8_e3m4 an 8-bit floating point representation: sign bit, 3 bit exponent (bias 3), 4 bit mantissa.

• IEEE 754-compliant, with NaN and +/-inf.

Float8E4M3

float8_e4m3 an 8-bit floating point representation: sign bit, 4 bit exponent (bias 7), 3 bit mantissa.

• IEEE 754-compliant, with NaN and +/-inf.

Float8E4M3B11FNUZ

float8_e4m3b11fnuz an 8-bit floating point representation: sign bit, 4 bit exponent (bias 11), 3 bit mantissa.

• Extended range: no infinity, NaN represented by 0b1000'0000.
• Subnormal numbers when biased exponent is 0.

Float8E4M3FNUZ

float8_e4m3fnuz an 8-bit floating point representation: sign bit, 4 bit exponent (bias 8), 3 bit mantissa.

• Extended range: no infinity, NaN represented by 0b1000'0000.
• Subnormal numbers when biased exponent is 0.

Float8E5M2

float8_e5m2 an 8-bit floating point representation: sign bit, 5 bit exponent (bias 15), 2 bit mantissa.

• IEEE 754-compliant, with NaN and +/-inf.
• Subnormal numbers when biased exponent is 0.

Float8E5M2FNUZ

float8_e5m2fnuz an 8-bit floating point representation: sign bit, 5 bit exponent (bias 16), 2 bit mantissa.

• Extended range: no infinity, NaN represented by 0b1000'0000.
• Subnormal numbers when biased exponent is 0.

Float8E8M0FNU

float8_e8m0fnu an 8-bit floating point representation: no sign bit, 8 bit exponent (bias 127), 0 bit mantissa.

• No zero, no infinity, NaN represented by 0b1111'1111.
• No subnormal numbers.

BFloat16

bfloat16 brain floating point data type: sign bit, 5 bits exponent, 10 bits mantissa.

Float16

float16 IEEE 754 half-precision floating point: sign bit, 5 bits exponent, 10 bits mantissa.

Float32

float32 IEEE 754 single-precision floating point: sign bit, 8 bits exponent, 23 bits mantissa.

Float64

float64 IEEE 754 double-precision floating point: sign bit, 11 bits exponent, 52 bits mantissa.

ComplexBFloat16

complex_float32 real and complex components are each brain floating point data type.

ComplexFloat16

complex_float32 real and complex components are each IEEE 754 half-precision floating point.

ComplexFloat32

complex_float32 real and complex components are each IEEE 754 single-precision floating point.

ComplexFloat64

complex_float64 real and complex components are each IEEE 754 double-precision floating point.

ComplexFloat4E2M1FN

complex_float4_e2m1fn real and complex components are each the float4_e2m1fn type.

ComplexFloat6E2M3FN

complex_float6_e2m3fn real and complex components are each the float6_e2m3fn type.

ComplexFloat6E3M2FN

complex_float6_e3m2fn real and complex components are each the float6_e3m2fn type.

ComplexFloat8E3M4

complex_float8_e3m4 real and complex components are each the float8_e3m4 type.

ComplexFloat8E4M3

complex_float8_e4m3 real and complex components are each the float8_e4m3 type.

ComplexFloat8E4M3B11FNUZ

complex_float8_e4m3b11fnuz real and complex components are each the float8_e4m3b11fnuz type.

ComplexFloat8E4M3FNUZ

complex_float8_e4m3fnuz real and complex components are each the float8_e4m3fnuz type.

ComplexFloat8E5M2

complex_float8_e5m2 real and complex components are each the float8_e5m2 type.

ComplexFloat8E5M2FNUZ

complex_float8_e5m2fnuz real and complex components are each the float8_e5m2fnuz type.

ComplexFloat8E8M0FNU

complex_float8_e8m0fnu real and complex components are each the float8_e8m0fnu type.

Complex64

complex64 real and complex components are each IEEE 754 single-precision floating point.

Complex128

complex128 real and complex components are each IEEE 754 double-precision floating point.

RawBits(usize)

r* raw bits, variable size given by *, limited to be a multiple of 8.

String

A UTF-8 encoded string.

Bytes

Variable-sized binary data.

NumpyDateTime64

numpy.datetime64 a 64-bit signed integer represents moments in time relative to the Unix epoch.

This data type closely models the datetime64 data type from NumPy.

Fields

unit: NumpyTimeUnit

The NumPy temporal unit.

scale_factor: NonZeroU32

The NumPy temporal scale factor.

NumpyTimeDelta64

numpy.timedelta64 a 64-bit signed integer represents signed temporal durations.

This data type closely models the timedelta64 data type from NumPy.

Fields

unit: NumpyTimeUnit

The NumPy temporal unit.

scale_factor: NonZeroU32

The NumPy temporal scale factor.

Extension(Arc<dyn DataTypeExtension>)

An extension data type.

Implementations

impl DataType

pub fn name(&self) -> String

Returns the name.

pub fn metadata(&self) -> MetadataV3

Returns the metadata.

pub fn size(&self) -> DataTypeSize

Returns the DataTypeSize.

pub fn fixed_size(&self) -> Option<usize>

Returns the size in bytes of a fixed-size data type, otherwise returns None.

pub fn from_metadata( metadata: &MetadataV3, data_type_aliases: &ExtensionAliasesDataTypeV3, ) -> Result<Self, PluginCreateError>

Create a data type from metadata.

Errors

Returns PluginCreateError if the metadata is invalid or not associated with a registered data type plugin.

pub fn fill_value_from_metadata( &self, fill_value: &FillValueMetadataV3, ) -> Result<FillValue, DataTypeFillValueMetadataError>

Create a fill value from metadata.

Errors

Returns DataTypeFillValueMetadataError if the fill value is incompatible with the data type.

pub fn metadata_fill_value( &self, fill_value: &FillValue, ) -> Result<FillValueMetadataV3, DataTypeFillValueError>

Create fill value metadata.

Errors

Returns an DataTypeFillValueError if the metadata cannot be created from the fill value.

Trait Implementations

impl Clone for DataType

fn clone(&self) -> DataType

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for DataType

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for DataType

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl PartialEq for DataType

fn eq(&self, other: &Self) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Eq for DataType