libcudf  24.04.00
Public Member Functions | List of all members
cudf::experimental::row::hash::row_hasher Class Reference

Computes the hash value of a row in the given table. More...

Public Member Functions

 row_hasher (table_view const &t, rmm::cuda_stream_view stream)
 Construct an owning object for hashing the rows of a table. More...
 
 row_hasher (std::shared_ptr< preprocessed_table > t)
 Construct an owning object for hashing the rows of a table from an existing preprocessed_table. More...
 
template<template< typename > class hash_function = cudf::hashing::detail::default_hash, template< template< typename > class, typename > class DeviceRowHasher = device_row_hasher, typename Nullate >
DeviceRowHasher< hash_function, Nullate > device_hasher (Nullate nullate={}, uint32_t seed=DEFAULT_HASH_SEED) const
 Get the hash operator to use on the device. More...
 

Detailed Description

Computes the hash value of a row in the given table.

Definition at line 1973 of file experimental/row_operators.cuh.

Constructor & Destructor Documentation

◆ row_hasher() [1/2]

cudf::experimental::row::hash::row_hasher::row_hasher ( table_view const &  t,
rmm::cuda_stream_view  stream 
)
inline

Construct an owning object for hashing the rows of a table.

Parameters
tThe table containing rows to hash
streamThe stream to construct this object on. Not the stream that will be used for comparisons using this object.

Definition at line 1982 of file experimental/row_operators.cuh.

◆ row_hasher() [2/2]

cudf::experimental::row::hash::row_hasher::row_hasher ( std::shared_ptr< preprocessed_table t)
inline

Construct an owning object for hashing the rows of a table from an existing preprocessed_table.

This constructor allows independently constructing a preprocessed_table and sharing it among multiple row_hasher and equality::self_comparator objects.

Parameters
tA table preprocessed for hashing or equality.

Definition at line 1996 of file experimental/row_operators.cuh.

Member Function Documentation

◆ device_hasher()

template<template< typename > class hash_function = cudf::hashing::detail::default_hash, template< template< typename > class, typename > class DeviceRowHasher = device_row_hasher, typename Nullate >
DeviceRowHasher<hash_function, Nullate> cudf::experimental::row::hash::row_hasher::device_hasher ( Nullate  nullate = {},
uint32_t  seed = DEFAULT_HASH_SEED 
) const
inline

Get the hash operator to use on the device.

Returns a unary callable, F, with signature hash_function::hash_value_type F(size_type).

F(i) returns the hash of row i.

Template Parameters
NullateA cudf::nullate type describing whether to check for nulls
Parameters
nullateIndicates if any input column contains nulls
seedThe seed to use for the hash function
Returns
A hash operator to use on the device

Definition at line 2014 of file experimental/row_operators.cuh.


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