Struct CompactTupleSketch

Source

pub struct CompactTupleSketch<S> { /* private fields */ }

Available on crate feature tuple and (crate features theta or tuple) only.

Expand description

Compact (immutable) Tuple sketch.

This is the serialization-friendly form: a compact array of retained hash-summary pairs plus theta and a 16-bit seed hash. It can be ordered (sorted ascending by hash) or unordered.

Implementations§

Source §

impl<S> CompactTupleSketch<S>

Source

pub fn as_view(&self) -> TupleSketchView<'_, S>

Returns a read-only view accepted by Tuple set operations.

Source

pub fn estimate(&self) -> f64

Returns the cardinality (distinct key count) estimate.

Source

pub fn theta(&self) -> f64

Returns theta as a fraction (0.0 to 1.0).

Source

pub fn theta64(&self) -> u64

Returns theta as u64.

Source

pub fn is_empty(&self) -> bool

Returns true if the sketch is empty.

Source

pub fn is_estimation_mode(&self) -> bool

Returns true if the sketch is in estimation mode.

Source

pub fn num_retained(&self) -> usize

Returns the number of retained entries.

Source

pub fn is_ordered(&self) -> bool

Returns true if retained entries are ordered (sorted ascending by hash).

Source

pub fn seed_hash(&self) -> u16

Returns the 16-bit seed hash.

Source

pub fn iter(&self) -> impl Iterator<Item = &TupleEntry<S>> + '_

Returns an iterator over retained entries.

Source

pub fn lower_bound(&self, num_std_dev: NumStdDev) -> f64

Returns the approximate lower error bound given the number of standard deviations.

Source

pub fn upper_bound(&self, num_std_dev: NumStdDev) -> f64

Returns the approximate upper error bound given the number of standard deviations.

Source

pub fn estimated_size(&self) -> usize

Returns the estimated size of the sketch in bytes.

Source

pub fn serialize(&self) -> Vec<u8> ⓘ
where S: TupleSummaryValue,

Serializes this sketch into the compact Tuple binary format.

Each summary is encoded by its TupleSummaryValue implementation. The layout matches the Java/C++ Tuple sketches, so the output can be read by those implementations given a compatible summary encoding.

§Examples

use datasketches::tuple::DefaultUpdatePolicy;
use datasketches::tuple::TupleSketchBuilder;

let policy = DefaultUpdatePolicy::<u64>::default();
let mut sketch = TupleSketchBuilder::new(policy).build().unwrap();
sketch.update("apple", 1);
let bytes = sketch.compact(true).serialize();
assert!(!bytes.is_empty());

Source

pub fn deserialize(bytes: &[u8]) -> Result<Self, Error>
where S: TupleSummaryValue,

Deserializes a compact Tuple sketch using the default seed.

§Errors

Returns InvalidData if the image is malformed, its seed hash does not match the default seed, or a summary cannot be decoded by S.

Source

pub fn deserialize_with_seed(bytes: &[u8], seed: u64) -> Result<Self, Error>
where S: TupleSummaryValue,

Deserializes a compact Tuple sketch using the provided expected seed.

§Errors

Returns InvalidData if the bytes are truncated, the family/serial version/sketch type are unexpected, the seed hash does not match, the supplied seed computes to the reserved zero seed hash, or an entry is corrupted.