pbmclapply {pbmcapply}

Tracking mclapply with progress bar

Description

pbmclapply is a wrapper around the mclapply function. It adds a progress bar to mclapply function.

Parallelization (mc.core > 1) works only on *nix (Linux, Unix such as macOS) system due to the lack of fork() functionality, which is essential for mcapply, on Windows.

Usage

pbmclapply(X, FUN, ...,
           mc.style = "ETA", mc.substyle = NA,
           mc.cores = getOption("mc.cores", 2L),
           ignore.interactive = getOption("ignore.interactive", F),
           mc.preschedule = TRUE, mc.set.seed = TRUE,
           mc.cleanup = TRUE, mc.allow.recursive = TRUE)

Arguments

X

a vector (atomic or list) or an expressions vector. Other objects (including classed objects) will be coerced by 'as.list'.

FUN

the function to be applied to.

...

optional arguments to FUN.

mc.cores

see mclapply.

mc.style, mc.substyle

style of the progress bar. See progressBar.

ignore.interactive

whether the interactive() is ignored. If set to TRUE, the progress bar will be printed even in a non-interactive environment (e.g. called by Rscript). Can be set as an option "ignore.interactive".

mc.preschedule, mc.set.seed, mc.cleanup, mc.allow.recursive

See mclapply.

Examples

# A lazy sqrt function which doesn't care about efficiency
lazySqrt <- function(num) {
  # Sleep randomly between 0 to 0.5 second
  Sys.sleep(runif(1, 0, 0.5))
  return(sqrt(num))
}

# On Windows, set cores to be 1
if (.Platform$OS.type == "windows") {
  cores = 1
} else {
  cores = 2
}

# A lazy and chatty sqrt function.
# An example of passing arguments to pbmclapply.
lazyChattySqrt <- function(num, name) {
  # Sleep randomly between 0 to 0.5 second
  Sys.sleep(runif(1, 0, 0.5))
  return(sprintf("Hello %s, the sqrt of %f is %f.", toString(name), num, sqrt(num)))
}

# Get the sqrt of 1-3 in parallel
result <- pbmclapply(1:3, lazySqrt, mc.cores = cores)
chattyResult <- pbmclapply(1:3, lazyChattySqrt, "Bob", mc.cores = cores)