mcsplitapply: Parallel split-matrix or dataframe apply loop

In traversc/trqwe: Performance oriented statistical metrics and utility functions.

Description

Splits a matrix or data.frame into subsets based on a factor, and applies a function to each subset. Typical use case: sum exon count data to gene count data. This is similar to the dplyr idiom: df %>% group_by(f) %>% do(...), but has several advantages. 1) mcsplitapply can be used on matrices (and is therefore much faster), 2) inherently parallelized, 3) can return results other than dataframes, 4) you can specify how the data are combined (default is rbind).

Usage

mcsplitapply(mat, f, func, mc.cores = 4, .combine = rbind, ...)

Arguments

mat	The matrix.
f	A factor of length equal to nrow(mat). The levels of this factor will split the matrix into subsets.
func	The function to apply to each subset.
mc.cores	The number of cores to use.
.combine	The function to combine the results with. Default is rbind. Use NA to return a list.

Value

A list or a combined object depending on the .combine parameter.

Examples

library(pasilla)
library(DEXSeq)
library(trqwe)
data(pasillaDEXSeqDataSet)
exon_counts <- counts(dxd)
f <- rowData(dxd)$groupID
gene_counts <- mcsplitapply(exon_counts, f, colSums)

traversc/trqwe documentation built on Dec. 4, 2020, 4:21 a.m.