Struct DataProfile 

pub struct DataProfile { /* private fields */ }

Sampling-based encoder selection

Strategy

1. Self::prune_candidates - “Prune”: Compute lightweight statistics over a representative sample of the data (average run length, sort order, max bit-width) and use them to prune obviously unsuitable candidates early.
2. Self::compete_u32 / Self::compete_u64 - “Compete”: Encode the same sample with every surviving candidate and pick the one whose encoded output is smallest. In case of a tie
   • the physical priority order is FastPFOR > VarInt > None and,
   • at the logical level, more complex transforms are deprioritized.

Implementations

impl DataProfile

pub fn prune_candidates<T>(values: &[T::UInt]) -> Vec<IntEncoder>

where T: ZigZag, <T as ZigZag>::UInt: WrappingSub,

Profile a representative sample to prune unsuitable candidates.

pub fn compete_u32(candidates: &[IntEncoder], data: &[u32]) -> IntEncoder

pub fn compete_u64(candidates: &[IntEncoder], data: &[u64]) -> IntEncoder

Trait Implementations

impl Clone for DataProfile

fn clone(&self) -> DataProfile

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for DataProfile

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

Formats the value using the given formatter.

impl Default for DataProfile

fn default() -> DataProfile

Returns the “default value” for a type.