SparseArray

SparseArray provides array-like containers for efficient in-memory representation of multidimensional sparse data in R, supporting both arrays and matrices with sparse-specific internal layouts.

Key Features:

• SparseArray Virtual Class: Defines a virtual class with two concrete subclasses, COO_SparseArray and SVT_SparseArray, for flexible sparse data representation.
• COO Layout: Implements coordinate list format as an internal representation for nonzero multidimensional data in COO_SparseArray objects.
• SVT Layout: Implements an alternative sparse vector tree format as an internal representation for nonzero multidimensional data in SVT_SparseArray objects.
• Base R Compatibility: SVT_SparseArray objects mimic ordinary matrix and array objects in base R, supporting most of the standard matrix and array API defined in base R.
• matrixStats Integration: Supports most of the standard matrix and array API defined in the matrixStats package from CRAN.

Scientific Applications:

• Multidimensional Sparse Data Storage: Enables efficient in-memory representation of sparse arrays and matrices commonly encountered in bioinformatics analyses.
• Sparse Matrix Operations: Supports computational workflows requiring sparse matrix manipulations while maintaining compatibility with existing R-based analytical pipelines.

Methodology:

The package implements two internal data layouts for sparse representation: the COO (coordinate list) layout and the SVT (sparse vector tree) layout, encapsulated within the COO_SparseArray and SVT_SparseArray concrete subclasses respectively.