MultiArray< N, T > Struct Template Reference [Multidimensional Arrays with Flexible Data Layout and On-Demand Slice Creation]

A multidimensional array for data that are stored via a native, writable C++ pointer. More...

#include <MultiArray.hpp>

List of all members.

Public Types

• typedef MultiArrayBase< N, T > Base_t
• typedef Base_t::value_type value_type

  The type of each element in this array.
  

• typedef MultiArray MultiArray_t
• typedef Base_t::ValueBase_t ValueBase_t
• typedef Base_t::component_t component_t
• typedef MultiArray< N,
  ValueBase_t > MultiArrayBase_t
• typedef MultiArray< N,
  component_t > MultiArrayComponent_t
• typedef Iterator< T > SliceStorage_t

  The layout of the stored data within a slice.
  

• typedef MultiArray< N-1, T >
  ::MultiArray_t Hyperslab_t

Public Member Functions

• MultiArray (T *data, const MultiIndex< N > &Sz)

  Construct a multidimensional array from a data pointer and size specification.
  

• MultiArray (const MultiIndex< N > &Sz, DataCreator< T > *Crec)

  Construct multidimensional array with callback for data creation when a certain slice is accessed (deferred data creation).
  

• MultiArray (const Iterator< T > &data, const MultiIndex< N > &Sz)

  Construct multidimensional array from iterator object and dimensionality.
  

• MultiArray (const MultiArray &MA)
• MultiArray (const Base_t &MA)
• MultiArrayBase_t arraybase () const
• operator MultiArrayBase_t () const

  Via implicit type conversion, any multidimensional array of a compound data type can be converted into a multidimensional array for a FixedArray<> instance.
  

• MultiArrayComponent_t component (int i) const
• Hyperslab_t operator[] (index_t i) const

  Extract an N-1 dimensional hyperslab of the current data set.
  

---

Detailed Description

template<int N, class T>
struct MultiArray< N, T >

A multidimensional array for data that are stored via a native, writable C++ pointer.

Note that no memory management of whatever kind is performed here, data allocation and deletion has to be done somewhere else.

Indexing by consecuted [] operators is REVERSED as to using a MultiIndex because the LAST index is running fastest!

Examples:

ColoredLines.cpp, CreatorExample.cpp, HelicalLines.cpp, InterpolatedDerivative.cpp, InterpolatedVectorDerivative.cpp, Interpolation.cpp, MonochromeSurface.cpp, NumericalGeodesics3D.cpp, NumericalGeodesics4D.cpp, NumericalPolarGeodesic2D.cpp, NumericalPolarSchwarzschild.cpp, PolychromeSurface.cpp, Sphere.cpp, and uniform_scalar.cpp.

---

Constructor & Destructor Documentation

template<int N, class T>

MultiArray< N, T >::MultiArray	(	const Iterator< T > &	data,
		const MultiIndex< N > &	Sz
	)			[inline]

Construct multidimensional array from iterator object and dimensionality.

No on-demand data creation here.

---

Member Function Documentation

template<int N, class T>

Hyperslab_t MultiArray< N, T >::operator[]

(

index_t

i

)

const [inline]

Extract an N-1 dimensional hyperslab of the current data set.

The returned objects contain a pointer to the original data (which are shared with the hyperslab) and N-1 indices referring to the respective dimensions. Note that the hyperslab data reside contigously in memory (but might still be a separated compound type, i.e. a compound type with its component distributed in different arrays).

Indexing via consecutive [] operators is SLOW as compared with multidimensional index access.

Reimplemented from MultiArrayBase< N, T >.