ReservoirSampling< Type, RNG > Class Template Reference

Detailed Description

template<typename Type, typename RNG = std::default_random_engine>
class thrill::common::ReservoirSampling< Type, RNG >

Implementation of reservoir sampling using Vitter's Algorithm R.

The reservoir size is fixed, new items replace old ones such that all items in the stream are sampled with the same uniform probability.

Definition at line 32 of file reservoir_sampling.hpp.

#include <reservoir_sampling.hpp>

Public Member Functions
	ReservoirSampling (size_t size, std::vector< Type > &samples, RNG &rng)
	initialize reservoir sampler More...
void	add (const Type &item)
	visit item, maybe add it to the sample. More...
size_t	count () const
	number of items seen More...
const std::vector< Type > &	samples () const
	access to samples More...
size_t	size () const
	size of reservoir More...

Private Attributes
size_t	count_ = 0
	number of items seen More...
RNG &	rng_
	source of randomness More...
std::vector< Type > &	samples_
	reservoir More...
size_t	size_
	size of reservoir More...

Constructor & Destructor Documentation

ReservoirSampling()

ReservoirSampling ( size_t  size, std::vector< Type > &  samples, RNG &  rng  )

inline

initialize reservoir sampler

Definition at line 36 of file reservoir_sampling.hpp.

References ReservoirSampling< Type, RNG >::samples_, and ReservoirSampling< Type, RNG >::size_.

Member Function Documentation

add()

void add ( const Type &  item )

inline

visit item, maybe add it to the sample.

Definition at line 43 of file reservoir_sampling.hpp.

References ReservoirSampling< Type, RNG >::count_, ReservoirSampling< Type, RNG >::rng_, ReservoirSampling< Type, RNG >::samples_, ReservoirSampling< Type, RNG >::size_, and gen_data::x.

count()

size_t count ( ) const

inline

number of items seen

Definition at line 61 of file reservoir_sampling.hpp.

References ReservoirSampling< Type, RNG >::count_.

samples()

const std::vector<Type>& samples ( ) const

inline

access to samples

Definition at line 64 of file reservoir_sampling.hpp.

References ReservoirSampling< Type, RNG >::samples_.

size()

size_t size ( ) const

inline

size of reservoir

Definition at line 58 of file reservoir_sampling.hpp.

References ReservoirSampling< Type, RNG >::size_.

Member Data Documentation

count_

size_t count_ = 0

private

number of items seen

Definition at line 70 of file reservoir_sampling.hpp.

Referenced by ReservoirSampling< Type, RNG >::add(), ReservoirSamplingFast< ValueType, decltype(rng_)>::add(), ReservoirSamplingGrow< SampleIndexPair >::add(), ReservoirSamplingGrow< SampleIndexPair >::calc_sample_size(), ReservoirSampling< Type, RNG >::count(), ReservoirSamplingFast< ValueType, decltype(rng_)>::count(), and ReservoirSamplingGrow< SampleIndexPair >::count().

rng_

RNG& rng_

private

source of randomness

Definition at line 74 of file reservoir_sampling.hpp.

Referenced by ReservoirSampling< Type, RNG >::add(), ReservoirSamplingFast< ValueType, decltype(rng_)>::add(), ReservoirSamplingGrow< SampleIndexPair >::add(), and ReservoirSamplingFast< ValueType, decltype(rng_)>::ReservoirSamplingFast().

samples_

std::vector<Type>& samples_

private

reservoir

Definition at line 72 of file reservoir_sampling.hpp.

Referenced by ReservoirSampling< Type, RNG >::add(), ReservoirSamplingFast< ValueType, decltype(rng_)>::add(), ReservoirSamplingGrow< SampleIndexPair >::add(), ReservoirSampling< Type, RNG >::ReservoirSampling(), ReservoirSamplingFast< ValueType, decltype(rng_)>::ReservoirSamplingFast(), ReservoirSampling< Type, RNG >::samples(), ReservoirSamplingFast< ValueType, decltype(rng_)>::samples(), and ReservoirSamplingGrow< SampleIndexPair >::samples().

size_

size_t size_

private

size of reservoir

Definition at line 68 of file reservoir_sampling.hpp.

Referenced by ReservoirSampling< Type, RNG >::add(), ReservoirSamplingFast< ValueType, decltype(rng_)>::add(), ReservoirSamplingGrow< SampleIndexPair >::add(), ReservoirSamplingGrow< SampleIndexPair >::calc_next_gap(), ReservoirSampling< Type, RNG >::ReservoirSampling(), ReservoirSamplingFast< ValueType, decltype(rng_)>::ReservoirSamplingFast(), ReservoirSampling< Type, RNG >::size(), ReservoirSamplingFast< ValueType, decltype(rng_)>::size(), and ReservoirSamplingGrow< SampleIndexPair >::size().

---

The documentation for this class was generated from the following file:

• thrill/common/reservoir_sampling.hpp