SparseCompressedColumnMatrix<T> Class

[SerializableAttribute]
public sealed class SparseCompressedColumnMatrix<T> : SparseMatrix<T>, 
	IResizableMatrix<T>, ISerializable

<SerializableAttribute>
Public NotInheritable Class SparseCompressedColumnMatrix(Of T)
	Inherits SparseMatrix(Of T)
	Implements IResizableMatrix(Of T), ISerializable

[SerializableAttribute]
generic<typename T>
public ref class SparseCompressedColumnMatrix sealed : public SparseMatrix<T>, 
	IResizableMatrix<T>, ISerializable

[<SealedAttribute>]
[<SerializableAttribute>]
type SparseCompressedColumnMatrix<'T> = 
    class
        inherit SparseMatrix<'T>
        interface IResizableMatrix<'T>
        interface ISerializable
    end

Inheritance

Object  →  LinearOperator<T>  →  Matrix<T>  →  SparseMatrix<T>  →  SparseCompressedColumnMatrix<T>

Implements

IResizableMatrix<T>, ISerializable

Type Parameters

T

Use the SparseCompressedColumnMatrix<T> class to represent sparse matrices. Only the nonzero elements of the matrix are stored. This not only saves memory, but also allows many calculations to be performed faster.

The compressed sparse column format stores the components in column-major order. Within each column, only the row indexes and the values of the nonzero components are stored.

The components of a sparse matrix can be specified in a variety of ways. A common method is to supply the rows, columns, and values of the nonzero components. Another way is to build up the matrix using the

Insert

Sparse matrices can also be read directly from values files in Matrix Market format.

The NonzeroCount property returns the number of nonzero components. The FillFactor property returns a number between 0 and 1 that is the proportion of nonzero components to the total number of components in the matrix. To iterate through the nonzero components of a sparse matrix, use the NonzeroElements property. This property enumerates RowColumnValueTriplet<T> objects, each containing the row, column, and value of a nonzero component.

A disadvantage of sparse matrices is that working with individual components tends to be slower. Every access usually requires a lookup in a list. To avoid duplicate lookups for elementary operations on components, use one of the specialized methods: AddAt(Int32, Int32, T), SubtractAt(Int32, Int32, T), MultiplyAt(Int32, Int32, T) and DivideAt(Int32, Int32, T).

Reference