Multi-dimensional array of type T. More...

#include <array_nd.hpp>

Inherits ivl::array_common_base< C >, and ivl::array_nd_details::basic_iterator_nd_interface< C >.

Public Types
typedef types::t_true	has_1d_parenthesis
	types that specify the defined parenthesis operators for this class
typedef const array< size_t, tiny > &	size_nd_ref_type
	return type of size_nd in that data-specific class
typedef array< size_t, tiny >	size_type
	parameter of resize, in each class of the hierarchy
typedef const array< size_t, tiny > &	stride_ref_type
	return type of size_nd in that data-specific class
Public Member Functions
base_class &	base ()
	get base class
Element access (multidimensional)
Elements can be accessed by specifying the
best_reference	operator() (size_t s1)
best_reference	operator() (size_t s1, size_t s2)
best_reference	operator() (size_t s1, size_t s2, size_t s3)
best_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4)
best_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4, size_t s5)
best_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4, size_t s5, size_t s6)
template<class S >
best_reference	operator() (const array< size_t, S > &indx)
const_reference	operator() (size_t s1) const
const_reference	operator() (size_t s1, size_t s2) const
const_reference	operator() (size_t s1, size_t s2, size_t s3) const
const_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4) const
const_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4, size_t s5) const
const_reference	operator() (size_t s1, size_t s2, size_t s3, size_t s4, size_t s5, size_t s6) const
template<class S >
const_reference	operator() (const array< size_t, S > &indx) const
Single-dimensional constructors
	array_nd ()
	Default constructor.
	array_nd (size_t count)
	Construct by defining an element count.
	array_nd (size_t count, const T &s)
	Construct by defining an element count and a sigle default value for all elements.
template<class J , class D >
	array_nd (size_t count, const array< J, D > &s)
	Construct by defining an element count and a sigle default value for all elements or an ivl::array to copy the elements from.
	array_nd (size_t count, const T *data)
	Construct by defining an element count and a pointer to data.
template<class S >
	array_nd (size_t count, const data_init_arg &y)
	Construct from a size_t and a data_init_arg used for arrays of special types only.
Multi-dimensional constructors
template<class S >
	array_nd (const array< size_t, S > &sz)
	Construct from a size_array containing the dimension sizes.
template<class S >
	array_nd (const array< size_t, S > &sz, const T &s)
	Construct from a size_array containing the dimension sizes and assign scalar.
template<class S >
	array_nd (const array< size_t, S > &sz, const T *ptr)
	Construct from a size_array containing the dimension sizes and assign values from an one-dimensional C array.
template<class S >
	array_nd (const array< size_t, S > &sz, const data_init_arg &y)
	Construct from a size_array and a data_init_arg used for arrays of special types only.
template<class S , class J , class D >
	array_nd (const array< size_t, S > &sz, const array< J, D > &a)
	Construct from a size_array containing the dimension sizes and an array with the values.
template<class S , class J , class D >
	array_nd (const array< size_t, S > &sz, array< J, D > &a)
	Construct from a size_array containing the dimension sizes and an array with the values.
	array_nd (const this_type &a)
	Copy constructor.
template<class J , class S >
	array_nd (const array_nd< J, S > &a)
	Constructor using another form of array_nd.
template<class J , class S >
	array_nd (array_nd< J, S > &a)
	(special for ref_iterator) Constructor using another form of array_nd
Size functions (multidimensional)
size_nd_ref_type	size_nd () const
	Get the size for each dimension.
size_t	size_nd (size_t d) const
	Get the size for a specific dimension.
stride_ref_type	stride () const
	Get the stride for each dimension.
size_t	stride (size_t dim) const
	Get the stride for a particular dimension.
size_type	size () const
	Get the size for each dimension.
size_t	size (size_t d) const
	Get the size for a specific dimension.
size_t	ndims () const
	Get the number of dimensions.
Resize Functions
void	init (const size_t newsize, const T &s)
void	init (const size_t newsize)
	init (1-dim) without padding
template<class S >
void	init (const array< size_t, S > &newsize, const T &s)
	init (n-dim)
template<class S >
void	init (const array< size_t, S > &newsize)
	init (n-dim) without padding
void	resize (const size_t newsize, const T &s)
void	resize (const size_t newsize)
	resize (1-dim) without padding
template<class S >
void	resize (const array< size_t, S > &newsize, const T &s)
	resize (n-dim)
template<class S >
void	resize (const array< size_t, S > &newsize)
	resize (n-dim) without padding
void	reshape (const size_t newsize, const T &s)
void	reshape (const size_t newsize)
	reshape (1-dim) without padding
template<class S >
void	reshape (const array< size_t, S > &newsize, const T &s)
	reshape (n-dim)
template<class S >
void	reshape (const array< size_t, S > &newsize)
	reshape (n-dim) without padding
Assignment Operators
template<class K >
derived_type &	operator= (const K &k)
this_type &	operator= (const this_type &in)
	C++ copy-constructor.

Detailed Description

template<class T, class OPTS = mem>
class ivl::array_nd< T, OPTS >

Multi-dimensional array of type T.

This structure is used to define a multidimensional array of type T. The number of dimensions can be specified when constructing an object and can be altered anytime by using methods like resize(const size_array &, const T).

The simpliest way to access an array_nd element is by using operator() and specifying the offset on each dimension, like this:

 int i = a(2, 3, 2)

An array_nd inherites all the functionality the array class provides, which means that it can also be indexed as an one-dimensional array using the array subscript operator: []. Keep in mind that in array_nd the elements are stored in column-major order like Fortran and Matlab and not like C. So, the offset is computed like this:

offset = row + column*NUMROWS

This is the normal specialization. It also applies to unspecialized data_class_sets, however they have to be resizable, creatable and writable.

Constructor & Destructor Documentation

template<class T, class OPTS = mem>

ivl::array_nd< T, OPTS >::array_nd	(	size_t	count,
		const T *	data
	)		`[inline]`

Construct by defining an element count and a pointer to data.

Warning:: The constructor will try to read count*size_of(T) bytes from the memory pointed by data, without performing any kind of check. The user is responsible of making sure the memory is accessible and properly initialized.

Member Function Documentation

template<class T, class D >

void array_nd< T, D >::init	(	const size_t	newsize,
		const T &	s
	)

init (1-dim) The array is resized according to the new shape and new length. The old data that fits into the new shape is kept from the portion of the same shape that was common to the new and the old shape. In other words, the array-level data is moved, so that the rectangle, cube, or n-dim shape of data remains intact, and it is optionally padded with new values if it grows.

template<class T, class OPTS = mem>

template<class S >

const_reference ivl::array_nd< T, OPTS >::operator() ( const array< size_t, S > & indx ) const [inline]

Selects a single, const element of the array, using a size_array for the coordinates.

template<class T, class OPTS = mem>

template<class S >

best_reference ivl::array_nd< T, OPTS >::operator() ( const array< size_t, S > & indx ) [inline]

Selects a single, mutable element of the array, using a size_array for the coordinates.

template<class T, class D >

void array_nd< T, D >::reshape	(	const size_t	newsize,
		const T &	s
	)

reshape (1-dim). resizes and keeps the underlying array contents. changes shape, and discards the old data shape. faster than resize. The array is resized. The data that fits in the new length is not affected. It will be the same as the old if viewed as an array, and it will be folded into the dimensions of the array_nd according to the new shape. Padding is applied at the end of the data.

template<class T, class D >

void array_nd< T, D >::resize	(	const size_t	newsize,
		const T &	s
	)

resize (1-dim) The array is resized according to the new shape and new length. The old data that fits into the new shape is kept from the portion of the same shape that was common to the new and the old shape. In other words, the array-level data is moved, so that the rectangle, cube, or n-dim shape of data remains intact, and it is optionally padded with new values if it grows.

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

ivl/details/array_nd/impl/specialization/array_nd_class.hpp
ivl/details/array_nd/impl/specialization/array_nd_2d_class.hpp
ivl/details/array_nd/impl/specialization/array_nd_image_class.hpp

Public Types

Public Member Functions

Detailed Description

template<class T, class OPTS = mem> class ivl::array_nd< T, OPTS >

Constructor & Destructor Documentation

Member Function Documentation

template<class T, class OPTS = mem>
class ivl::array_nd< T, OPTS >