Collapse separate simulation files into a single result

This function collects and aggregates the results from SimDesign's runSimulation into a single objects suitable for post-analyses, or combines all the saved results directories and combines them into one. This is useful when results are run piece-wise on one node (e.g., 500 replications in one batch, 500 again at a later date, though be careful about the set.seed use as the random numbers will tend to correlate the more it is used) or run independently across different nodes/computing cores (e.g., see runArraySimulation.

Usage

SimCollect(
  dir = NULL,
  files = NULL,
  filename = NULL,
  select = NULL,
  check.only = FALSE,
  target.reps = NULL,
  warning_details = FALSE,
  error_details = TRUE,
  gc = FALSE
)

aggregate_simulations(...)

Arguments

dir: a character vector pointing to the directory name containing the .rds files. All .rds files in this directory will be used after first checking their status with SimCheck. For greater specificity use the files argument
files: a character vector containing the names of the simulation's final .rds files.
filename: (optional) name of .rds file to save aggregate simulation file to. If not specified then the results will only be returned in the R console.
select: a character vector indicating columns to variables to select from the SimExtract(what='results') information. This is mainly useful when RAM is an issue given simulations with many stored estimates. Default includes the results objects in their entirety, though to omit all internally stored simulation results pass the character 'NONE'. To investigate the stored warnings and error messages in isolation pass 'WARNINGS' or 'ERRORS', respectively
check.only: logical; for larger simulations file sets, such as those generated by runArraySimulation, return the design conditions that do no satisfy the target.reps and throw warning if files are unexpectedly missing
target.reps: (optional) number of replications to check against to evaluate whether the simulation files returned the desired number of replications. If missing, the highest detected value from the collected set of replication information will be used
warning_details: logical; include the aggregate of the warnings to be extracted via SimExtract?
error_details: logical; include the aggregate of the errors to be extracted via SimExtract?
gc: logical; explicitly call R's garbage collector gc? May help when memory is severely constrained during the file read-ins. Otherwise, the select argument should be used to take more memory-friendly subsets
...: not used

Value

returns a data.frame/tibble with the (weighted) average/aggregate of the simulation results

References

Chalmers, R. P., & Adkins, M. C. (2020). Writing Effective and Reliable Monte Carlo Simulations with the SimDesign Package. The Quantitative Methods for Psychology, 16(4), 248-280. doi:10.20982/tqmp.16.4.p248

Sigal, M. J., & Chalmers, R. P. (2016). Play it again: Teaching statistics with Monte Carlo simulation. Journal of Statistics Education, 24(3), 136-156. doi:10.1080/10691898.2016.1246953

Author

Phil Chalmers rphilip.chalmers@gmail.com

Examples

if (FALSE) { # \dontrun{

setwd('my_working_directory')

## run simulations to save the .rds files (or move them to the working directory)
# seeds1 <- genSeeds(design)
# seeds2 <- genSeeds(design, old.seeds=seeds1)
# ret1 <- runSimulation(design, ..., seed=seeds1, filename='file1')
# ret2 <- runSimulation(design, ..., seed=seeds2, filename='file2')

# saves to the hard-drive and stores in workspace
final <- SimCollect(files = c('file1.rds', 'file2.rds'))
final

# If filename not included, can be extracted from results
# files <- c(SimExtract(ret1, 'filename'), SimExtract(ret2, 'filename'))
# final <- SimCollect(files = files)


#################################################
# Example where each row condition is repeated, evaluated independently,
# and later collapsed into a single analysis object

# Each condition repeated four times (hence, replications
# should be set to desired.reps/4)
Design <- createDesign(mu = c(0,5),
                       N  = c(30, 60))
Design

# assume the N=60 takes longer, and should be spread out across more arrays
Design_long <- expandDesign(Design, c(2,2,4,4))
Design_long

replications <- c(rep(50, 4), rep(25,8))
data.frame(Design_long, replications)

#-------------------------------------------------------------------

Generate <- function(condition, fixed_objects) {
    dat <- with(condition, rnorm(N, mean=mu))
    dat
}

Analyse <- function(condition, dat, fixed_objects) {
    ret <- c(mean=mean(dat), SD=sd(dat))
    ret
}

Summarise <- function(condition, results, fixed_objects) {
    ret <- colMeans(results)
    ret
}

#-------------------------------------------------------------------

# create directory to store all final simulation files
dir.create('sim_files/')

iseed <- genSeeds()

# distribute jobs independently
sapply(1:nrow(Design_long), \(i) {
  runArraySimulation(design=Design_long, replications=replications,
                generate=Generate, analyse=Analyse, summarise=Summarise,
                arrayID=i, dirname='sim_files/', filename='job', iseed=iseed)
}) |> invisible()

# check that all replications satisfy target
SimCollect('sim_files/', check.only = TRUE)

# specify files explicitly
SimCollect(files = list.files(path='sim_files/', pattern="*.rds", full.names=TRUE),
   check.only = TRUE)

# this would have been returned were the target.rep supposed to be 1000
SimCollect('sim_files/', check.only = TRUE, target.reps=1000)

# aggregate into single object
sim <- SimCollect('sim_files/')
sim

# view list of error messages (if there were any raised)
SimCollect('sim_files/', select = 'ERRORS')

SimClean(dir='sim_files/')

} # }