Internal data structured used by XGBoost during training. More...

#include <data.h>

Public Member Functions
	DMatrix ()=default
	default constructor

virtual MetaInfo &	Info ()=0
	meta information of the dataset

virtual void	SetInfo (const char key, const void dptr, DataType dtype, size_t num)

virtual void	SetInfo (const char *key, std::string const &interface_str)

virtual const MetaInfo &	Info () const =0
	meta information of the dataset

XGBAPIThreadLocalEntry &	GetThreadLocal () const
	Get thread local memory for returning data from DMatrix.

virtual Context const *	Ctx () const =0
	Get the context object of this DMatrix.

template<typename T >
BatchSet< T >	GetBatches ()
	Gets batches.

template<typename T >
BatchSet< T >	GetBatches (Context const *ctx)

template<typename T >
BatchSet< T >	GetBatches (Context const *ctx, const BatchParam &param)

template<typename T >
bool	PageExists () const

virtual bool	SingleColBlock () const =0

virtual	~DMatrix ()
	virtual destructor

bool	IsDense () const
	Whether the matrix is dense.

virtual DMatrix *	Slice (common::Span< int32_t const > ridxs)=0

virtual DMatrix *	SliceCol (int num_slices, int slice_id)=0
	Slice a DMatrix by columns.

template<>
BatchSet< SparsePage >	GetBatches ()

template<>
bool	PageExists () const

template<>
bool	PageExists () const

template<>
bool	PageExists () const

template<>
BatchSet< SparsePage >	GetBatches (Context const *)

template<>
BatchSet< CSCPage >	GetBatches (Context const *ctx)

template<>
BatchSet< SortedCSCPage >	GetBatches (Context const *ctx)

template<>
BatchSet< EllpackPage >	GetBatches (Context const *ctx, BatchParam const &param)

template<>
BatchSet< GHistIndexMatrix >	GetBatches (Context const *ctx, BatchParam const &param)

template<>
BatchSet< ExtSparsePage >	GetBatches (Context const *ctx, BatchParam const &param)

Static Public Member Functions
static DMatrix *	Load (const std::string &uri, bool silent=true, DataSplitMode data_split_mode=DataSplitMode::kRow)
	Load DMatrix from URI.

template<typename AdapterT >
static DMatrix *	Create (AdapterT *adapter, float missing, int nthread, const std::string &cache_prefix="", DataSplitMode data_split_mode=DataSplitMode::kRow)
	Creates a new DMatrix from an external data adapter.

template<typename DataIterHandle , typename DMatrixHandle , typename DataIterResetCallback , typename XGDMatrixCallbackNext >
static DMatrix *	Create (DataIterHandle iter, DMatrixHandle proxy, std::shared_ptr< DMatrix > ref, DataIterResetCallback reset, XGDMatrixCallbackNext next, float missing, int nthread, bst_bin_t max_bin)
	Create a new Quantile based DMatrix used for histogram based algorithm.

template<typename DataIterHandle , typename DMatrixHandle , typename DataIterResetCallback , typename XGDMatrixCallbackNext >
static DMatrix *	Create (DataIterHandle iter, DMatrixHandle proxy, DataIterResetCallback reset, XGDMatrixCallbackNext next, float missing, int32_t nthread, std::string cache)
	Create an external memory DMatrix with callbacks.

Protected Member Functions
virtual BatchSet< SparsePage >	GetRowBatches ()=0

virtual BatchSet< CSCPage >	GetColumnBatches (Context const *ctx)=0

virtual BatchSet< SortedCSCPage >	GetSortedColumnBatches (Context const *ctx)=0

virtual BatchSet< EllpackPage >	GetEllpackBatches (Context const *ctx, BatchParam const &param)=0

virtual BatchSet< GHistIndexMatrix >	GetGradientIndex (Context const *ctx, BatchParam const &param)=0

virtual BatchSet< ExtSparsePage >	GetExtBatches (Context const *ctx, BatchParam const &param)=0

virtual bool	EllpackExists () const =0

virtual bool	GHistIndexExists () const =0

virtual bool	SparsePageExists () const =0

Detailed Description

Internal data structured used by XGBoost during training.

Member Function Documentation

◆ Create() [1/3]

template<typename AdapterT >

DMatrix * xgboost::DMatrix::Create	(	AdapterT *	adapter,
		float	missing,
		int	nthread,
		const std::string &	cache_prefix = `""`,
		DataSplitMode	data_split_mode = `DataSplitMode::kRow`
	)

static

Creates a new DMatrix from an external data adapter.

Template Parameters

AdapterT Type of the adapter.

Parameters

[in,out]	adapter	View onto an external data.
	missing	Values to count as missing.
	nthread	Number of threads for construction.
	cache_prefix	(Optional) The cache prefix for external memory.
	data_split_mode	(Optional) Data split mode.

Returns: a Created DMatrix.

◆ Create() [2/3]

template<typename DataIterHandle , typename DMatrixHandle , typename DataIterResetCallback , typename XGDMatrixCallbackNext >

DMatrix * xgboost::DMatrix::Create	(	DataIterHandle	iter,
		DMatrixHandle	proxy,
		DataIterResetCallback *	reset,
		XGDMatrixCallbackNext *	next,
		float	missing,
		int32_t	nthread,
		std::string	cache
	)

static

Create an external memory DMatrix with callbacks.

Template Parameters

DataIterHandle	External iterator type, defined in C API.
DMatrixHandle	DMatrix handle, defined in C API.
DataIterResetCallback	Callback for reset, prototype defined in C API.
XGDMatrixCallbackNext	Callback for next, prototype defined in C API.

Parameters

iter	External data iterator
proxy	A hanlde to ProxyDMatrix
reset	Callback for reset
next	Callback for next
missing	Value that should be treated as missing.
nthread	number of threads used for initialization.
cache	Prefix of cache file path.

Returns: A created external memory DMatrix.

◆ Create() [3/3]

template<typename DataIterHandle , typename DMatrixHandle , typename DataIterResetCallback , typename XGDMatrixCallbackNext >

DMatrix * xgboost::DMatrix::Create	(	DataIterHandle	iter,
		DMatrixHandle	proxy,
		std::shared_ptr< DMatrix >	ref,
		DataIterResetCallback *	reset,
		XGDMatrixCallbackNext *	next,
		float	missing,
		int	nthread,
		bst_bin_t	max_bin
	)

static

Create a new Quantile based DMatrix used for histogram based algorithm.

Template Parameters

DataIterHandle	External iterator type, defined in C API.
DMatrixHandle	DMatrix handle, defined in C API.
DataIterResetCallback	Callback for reset, prototype defined in C API.
XGDMatrixCallbackNext	Callback for next, prototype defined in C API.

Parameters

iter	External data iterator
proxy	A hanlde to ProxyDMatrix
ref	Reference Quantile DMatrix.
reset	Callback for reset
next	Callback for next
missing	Value that should be treated as missing.
nthread	number of threads used for initialization.
max_bin	Maximum number of bins.

Returns: A created quantile based DMatrix.

◆ Ctx()

virtual Context const * xgboost::DMatrix::Ctx ( ) const

pure virtual

Get the context object of this DMatrix.

The context is created during construction of DMatrix with user specified nthread parameter.

◆ GetBatches()

template<typename T >

BatchSet< T > xgboost::DMatrix::GetBatches ( )

Gets batches.

Use range based for loop over BatchSet to access individual batches.

◆ Load()

DMatrix * xgboost::DMatrix::Load	(	const std::string &	uri,
		bool	silent = `true`,
		DataSplitMode	data_split_mode = `DataSplitMode::kRow`
	)

static

Load DMatrix from URI.

Parameters

uri	The URI of input.
silent	Whether print information during loading.
data_split_mode	In distributed mode, split the input according this mode; otherwise, it's just an indicator on how the input was split beforehand.

Returns: The created DMatrix.

◆ SingleColBlock()

virtual bool xgboost::DMatrix::SingleColBlock ( ) const

pure virtual

Returns: Whether the data columns single column block.

◆ SliceCol()

virtual DMatrix * xgboost::DMatrix::SliceCol	(	int	num_slices,
		int	slice_id
	)

pure virtual

Slice a DMatrix by columns.

Parameters

num_slices	Total number of slices
slice_id	Index of the current slice

Returns: DMatrix containing the slice of columns

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

External/xgboost/include/xgboost/data.h
External/xgboost/src/data/data.cc

Public Member Functions

Static Public Member Functions

Protected Member Functions

Detailed Description

Member Function Documentation

◆ Create() [1/3]

◆ Create() [2/3]

◆ Create() [3/3]

◆ Ctx()

◆ GetBatches()

◆ Load()

◆ SingleColBlock()

◆ SliceCol()