fapply: Apply a function over a list or vector with optional...

In functionals: Functional Programming with Parallelism and Progress Tracking

Apply a function over a list or vector with optional parallelism and progress

Description

A lightweight and fast version of 'lapply()' with support for multicore (Unix) and snow-style clusters via 'parallel', with internal progress bar tracking and message suppression. When 'pb = TRUE', progress advances when individual tasks complete, including in the parallel paths, so counts reflect completed jobs rather than internal chunk boundaries. Progress rendering is throttled for large workloads to keep console redraw overhead bounded, while status output reports elapsed time and estimated time remaining.

Usage

fapply(.x, .f, ncores = 1, pb = FALSE, cl = NULL, load_balancing = TRUE, ...)

Arguments

.x	A list or atomic vector.
.f	Function to apply.
ncores	Number of cores to use (default: 1 = sequential).
pb	Show progress bar? (default: FALSE). When enabled, progress is driven by completed tasks. In parallel mode, updates occur as workers return results.
cl	A cluster object (from parallel::makeCluster), or integer for core count.
load_balancing	Logical. Use 'parLapplyLB' if 'TRUE' (default: 'FALSE').
...	Additional arguments passed to '.f'.

Value

A list of results.

Examples

# Basic usage (sequential)
fapply(1:5, sqrt)

# With progress bar (sequential)
fapply(1:5, function(x) { Sys.sleep(0.1); x^2 }, pb = TRUE)

# Multicore on Unix (if available)

if (.Platform$OS.type != "windows") {
  fapply(1:10, sqrt, ncores = 2)
}


# With user-created cluster (portable across platforms)

cl <- parallel::makeCluster(2)
fapply(1:10, sqrt, cl = cl)
parallel::stopCluster(cl)


# Heavy computation example with exact completion-driven progress

heavy_fn <- function(x) { Sys.sleep(0.05); x^2 }
fapply(1:20, heavy_fn, ncores = 2, pb = TRUE)

functionals documentation built on Aug. 22, 2026, 9:07 a.m.