Rscript

exec.r

#!/usr/bin/Rscript

library(RPostgreSQL)
library(data.table)
library(optparse)

opts.list <- list(
  make_option(c('-d', '--day'), help='Global Rank for day')
  # ... further options
)
opts <- OptionParser(option_list=opts.list)

source('db_connect.r')

tryCatch({
    # connect to database, redirect output streams, etc

    source('global_rank.r')
    calculate_global_rank(opts$options$day)
  }
  , finally={
    # close connections
  }
})

parallel.r

# this is part of R, not an external package
library(parallel)

# optional: impose reproducibility of the uniform random draws below
set.seed(11)

# take advantage of all cores on my computer
nodes = 8

# wrap longer tasks in a function
compute <- function(arg) {
  computing.times <- runif(1) * arg
  Sys.sleep(computing.times) # compute heavily!
  computing.times
}

# start an 8 workers parallel cluster
cluster <- makeCluster(nodes)

print(system.time({
  tryCatch(
    {
      args.range <- 1:nodes
      print(parLapply(cluster, args.range, compute))
    },
    finally=stopCluster(cluster)
  )
}))