Run jobs in other R sessions without waiting

This function has been deprecated. Please use lapply_callr instead.

Usage

async_works(
  X,
  FUN,
  ...,
  .globals = NULL,
  .name = "Untitled",
  .rs = FALSE,
  .wait = TRUE,
  .chunk_size = Inf,
  .nworkers = future::availableCores(),
  .simplify = FALSE,
  .quiet = FALSE,
  .log
)

Arguments

X

vector or list to be applied

FUN

function with the first argument to be each element of X

...

further arguments to be passed to FUN

.globals

global variables to be evaluated in FUN

.name

job names, used if backed by rstudioapi jobs

.rs

whether to use rstudioapi jobs

.wait

whether to wait for the results

.chunk_size

used only when .wait=FALSE, chunk size for each workers at a time. Only useful for printing progress messages, but might slow down the process when .chunk_size is too small

.nworkers

number of workers at a time

.simplify

whether to simplify the results, i.e. merge list of results to vectors or arrays

.quiet

whether to suppress the printing messages

.log

internally used

Value

If .wait=TRUE, returns the applied results of FUN on each of X. The result types depend on .simplify (compare the difference between lapply and sapply). If .wait=FALSE, then returns a function that can check the result. The function takes timeout argument that blocks the session at most timeout seconds waiting for the results. See examples.

Details

Unlike future package, where the global variables can be automatically detected, async_works require users to specify global variables explicitly via .globals

async_works is almost surely slower than future.apply packages. However, it provides a functionality that future.apply can hardly achieve: being non-block. When setting .wait=FALSE, the process will run in the background, and one may run as many of these tasks as they want. This is especially useful when large data generating process occurs ( such as read in from a file, process, generate summarizing reports).

Examples

if (FALSE) { # \dontrun{
# requires a sub-process to run the code

# Basic usage
a <- 1
async_works(1:10, function(ii){
  ii + a # sub-process don't know a, hence must pass a as globals
}, .globals = list(a = a))

# non-blocking case
system.time({
  check <- async_works(1:10, function(ii){
    # simulating process, run run run
    Sys.sleep(ii)
    Sys.getpid()
  }, .wait = FALSE)
})

# check the results
res <- check(timeout = 0.1)
attr(res, 'resolved') # whether it's resolved

# block the session waiting for the results
res <- check(timeout = Inf)
attr(res, 'resolved')


} # }