Matrix.to.Pointers: Extract row and column indices, pointers and values from a...

In sparseHessianFD: Numerical Estimation of Sparse Hessians

Extract row and column indices, pointers and values from a sparse matrix.

Description

Returns a list of row indices, column indices, pointers, and/or values of a sparse Hessian.

Usage

Matrix.to.Pointers(
  M,
  as.symmetric = Matrix::isSymmetric(M),
  values = !is(M, "nMatrix"),
  order = NULL,
  index1 = TRUE
)

Arguments

M	A sparse Matrix, as defined in the Matrix package.
as.symmetric	Defaults to isSymmetric(M). If M is symmetric, and as.symmetric is FALSE, then index/pointer elements in the output list will be labeled according to order. If M is not symmetric, and as.symmetric is TRUE, then an error will be triggered.
values	If TRUE, values are returned in list as 'x'. Defaults to TRUE for numeric and logical matrices, and FALSE for pattern matrices. If M is a pattern matrix, values=TRUE will trigger a warning.
order	Determines the indexing/compression scheme for the output matrix. Use 'triplet" to get row and column indices. Defaults to the same class as M.
index1	TRUE (default) if return indices and pointers should use 1-based indexing. FALSE for 0-based indexing.

Details

This function is included primarily for debugging purposes. It is used internally, but would not ordinarily be called by an end user.

Value

A list with the following elements. If order=='row',

jCol

Integer vector containing column indices of non-zero elements

ipntr

Integer vector containing pointers to elements of jCol at which the next row begins.

If order=='column'

iRow

Integer vector containing row indices of non-zero elements

jpntr

Integer vector containing pointers to elements of iRow at which the next column begins.

If order=='triplet'

rows

Row indices of non-zero elements

cols

Column indices of non-zero elements

If as.symmetric is TRUE, then the row/column orientation does not matter.

idx

Integer vector containing indices of non-zero elements

pntr

Integer vector containing pointers to elements of idx at which the next row or column begins.

If values=TRUE, the return list includes x, the values of the non-zero elements. The 'class' element is the name of the sparse matrix class to which the output corresponds (identifies numeric type, pattern, and indexing/compression scheme).

See Also

Matrix.to.Coord

sparseHessianFD documentation built on Oct. 20, 2022, 1:08 a.m.