Trait Encode 

pub trait Encode: Clone {
    const SIZE: DataSize;
    const ALIGNMENT: PageOffset;

    // Required methods
    fn encode(&self) -> Cow<'_, [u8]>;
    fn decode(data: Cow<'_, [u8]>) -> MemoryResult<Self>
       where Self: Sized;
    fn size(&self) -> MSize;
}

This trait defines the encoding and decoding behaviour for data types used in the DBMS.

Required Associated Constants

const SIZE: DataSize

The size characteristic of the data type.

The DataSize can either be a fixed size in bytes or dynamic.

const ALIGNMENT: PageOffset

The alignment requirement in bytes for the data type.

If Self::SIZE is DataSize::Fixed, the alignment must be equal to the size, otherwise it can be any value.

This value should never be less than 8 for DataSize::Dynamic types to ensure proper memory alignment.

We should set a default value (probably 32) for dynamic types to avoid misalignment issues, but letting an expert user to override it if necessary.

Required Methods

fn encode(&self) -> Cow<'_, [u8]>

Encodes the data type into a vector of bytes.

fn decode(data: Cow<'_, [u8]>) -> MemoryResult<Self>

where Self: Sized,

Decodes the data type from a slice of bytes.

fn size(&self) -> MSize

Returns the size in bytes of the encoded data type.

Implementations must never panic or wrap: when the encoded size does not fit an MSize, they must return MSize::MAX. No record of that size fits a page together with its header, so storage rejects it with MemoryError::DataTooLarge before calling Self::encode. The output of Self::encode is unspecified for such values.

Dyn Compatibility

This trait is not dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementations on Foreign Types

impl Encode for Vec<Value>

Encodes a Vec<Value> as [count: u32 LE] + [for each value: [size: u32 LE] + value.encode()].

This is used as the key type for B-tree indexes, supporting both single-column and composite indexes uniformly.

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT

fn encode(&self) -> Cow<'_, [u8]>

fn decode(data: Cow<'_, [u8]>) -> MemoryResult<Self>

fn size(&self) -> MSize

Implementors

impl Encode for Blob

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT

impl Encode for Boolean

const SIZE: DataSize

const ALIGNMENT: PageOffset = 1

impl Encode for ColumnSnapshot

const ALIGNMENT: PageOffset = 32

const SIZE: DataSize = DataSize::Dynamic

impl Encode for DataTypeSnapshot

const ALIGNMENT: PageOffset = 32

const SIZE: DataSize = DataSize::Dynamic

impl Encode for Date

const SIZE: DataSize

const ALIGNMENT: PageOffset = 4

impl Encode for DateTime

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Decimal

const SIZE: DataSize

const ALIGNMENT: PageOffset = RUST_DECIMAL_ENCODE_SIZE

impl Encode for ForeignKeySnapshot

const ALIGNMENT: PageOffset = 32

const SIZE: DataSize = DataSize::Dynamic

impl Encode for IndexSnapshot

const ALIGNMENT: PageOffset = 32

const SIZE: DataSize = DataSize::Dynamic

impl Encode for Int8

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Int16

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Int32

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Int64

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Json

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT

impl Encode for TableSchemaSnapshot

const ALIGNMENT: PageOffset = 32

const SIZE: DataSize = DataSize::Dynamic

impl Encode for Text

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT

impl Encode for Uint8

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Uint16

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Uint32

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Uint64

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Uuid

const SIZE: DataSize

const ALIGNMENT: PageOffset

impl Encode for Value

Encodes a Value as [discriminant: u8] + [inner_type.encode()].

For Null, only the discriminant byte is written. For Custom, the encoding is [discriminant] + [tag_len: u16 LE] + [tag_bytes] + [data_len: u16 LE] + [encoded_bytes].

Encode::size saturates at MSize::MAX when a value does not fit the 2-byte length prefixes; storage rejects such values before encoding.

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT

impl<T> Encode for Nullable<T>

where T: DataType,

const SIZE: DataSize = DataSize::Dynamic

const ALIGNMENT: PageOffset = DEFAULT_ALIGNMENT