solve.HDF5Matrix: Matrix inverse of a symmetric positive-definite HDF5Matrix...

In BigDataStatMeth: Title: Scalable Statistical Computing with HDF5-Backed Matrices

Matrix inverse of a symmetric positive-definite HDF5Matrix via Cholesky

Description

Computes the matrix inverse of a symmetric positive-definite HDF5Matrix using Cholesky decomposition + back-substitution. Equivalent to base::solve(A) for SPD matrices.

Usage

## S3 method for class 'HDF5Matrix'
solve(
  a,
  b,
  full_matrix = TRUE,
  overwrite = FALSE,
  threads = -1L,
  block_size = NULL,
  compression = NULL,
  ...
)

Arguments

a	An HDF5Matrix (square, symmetric positive-definite).
b	Not supported for HDF5Matrix; must be missing.
full_matrix	Logical. Return full symmetric inverse. Default TRUE.
overwrite	Logical. Overwrite existing result. Default FALSE.
threads	Integer. OpenMP threads (-1 = auto).
block_size	Integer or NULL. Elements per block. NULL = auto.
compression	Integer (0-9) or NULL. gzip compression level for the result dataset. NULL uses the global option set by hdf5matrix_options (default 6). Use 0 to disable compression (faster for benchmarks).
...	Ignored (for S3 compatibility).

Value

HDF5Matrix containing the matrix inverse.

Examples

tmp <- tempfile(fileext = ".h5")

X  <- hdf5_create_matrix(tmp, "data/X", data = matrix(rnorm(10000), 100, 100))

X   <- hdf5_matrix(tmp, "data/X")
AtA <- crossprod(X)              # HDF5Matrix, square SPD
inv <- solve(AtA)               # inverse of AtA

hdf5_close_all()
unlink(tmp)

BigDataStatMeth documentation built on May 15, 2026, 1:07 a.m.