Consider a temporary directory with four txt files

path <- file.path(tempdir(), "demo")
unlink(path)
dir.create(path)

writeLines("the contents of file.txt", file.path(path, "file.txt"))
writeLines("the contents of file2.txt", file.path(path, "file2.txt"))
writeLines("the contents of file3.txt", file.path(path, "file3.txt"))
writeLines("the contents of file4.txt", file.path(path, "file4.txt"))

and a function that will successfully process the first two files (in this case rewriting their contents) but fail for the last two by returning FALSE and throwing an error, respectively.

fun <- function(file) {
  if(grepl("file3[.]txt$", file)) return(FALSE)
  if(grepl("file4[.]txt$", file)) stop("Uh, Houston, we've had a problem.", call. = FALSE)
  txt <- readLines(file)
  txt <- gsub("contents", "modified contents", txt)
  writeLines(txt, file)
}

Configure

Now let the user configure the directory to only process files names that include a digit before the file extension.

library(batchr)
print(batch_config(fun, path = path, regexp = "file\\d[.]txt$"))
#> [1] "file2.txt" "file3.txt" "file4.txt"

The contents of the hidden figuration file are as follows

batch_config_read(path)
#> $time
#> [1] "2021-09-20 15:42:22 UTC"
#> 
#> $regexp
#> [1] "file\\d[.]txt$"
#> 
#> $recurse
#> [1] FALSE
#> 
#> $fun
#> function(file) {
#>   if(grepl("file3[.]txt$", file)) return(FALSE)
#>   if(grepl("file4[.]txt$", file)) stop("Uh, Houston, we've had a problem.", call. = FALSE)
#>   txt <- readLines(file)
#>   txt <- gsub("contents", "modified contents", txt)
#>   writeLines(txt, file)
#> }
#> 
#> $dots
#> list()

The time value specifies the system time (in UTC) that the project was configured. It is very important because only files that were last modified before the configuration time are considered to be unprocessed (when a file is successfully processed its modification time is automatically set to the current system time).

Run

With the directory configured the next task is to start processing the files

print(batch_run(path, ask = FALSE))
#> file2.txt file3.txt file4.txt 
#>      TRUE     FALSE     FALSE

From the output we can see that ‘file2.csv’ was processed successfully but processing of the other two files that matched regexp failed. The output is also recorded in a hidden log file that can be read using batch_log_read().

batch_log_read(path)
#>      type         time      file                           message
#> 1 SUCCESS 00:00:00.001 file2.txt                              <NA>
#> 2 FAILURE 00:00:00.007 file3.txt                              <NA>
#> 3 FAILURE 00:00:00.001 file4.txt Uh, Houston, we've had a problem.

or summarised using

batch_report(path)
#> ✔ file2.txt [00:00:00.001]
#> ✖ file3.txt [00:00:00.007]
#> ✖ file4.txt [00:00:00.001]
#> Success: 1
#> Failure: 2
#> Remaining: 0
#> 

The contents of ‘file2.csv’ are now as follows

readLines(file.path(path, "file2.txt"))
#> [1] "the modified contents of file2.txt"

Reconfigure

At this point let us update the function and regular expression so that all files are included and successfully processed.

batch_reconfig_fileset(path, regexp = "[.]txt$")

fun <- function(file) {
  txt <- readLines(file)
  txt <- gsub("contents", "modified contents", txt)
  writeLines(txt, file)
}

batch_reconfig_fun(path, fun)

Rerun

Now when we call batch_run() the newly included ‘file.txt’ is successfully processed.

batch_run(path, ask = FALSE)
batch_report(path)
#> ✔ file.txt [00:00:00.001]
#> ✔ file2.txt [00:00:00.001]
#> ✖ file3.txt [00:00:00.007]
#> ✖ file4.txt [00:00:00.001]
#> Success: 2
#> Failure: 2
#> Remaining: 0
#> 

In order to reattempt processing of the ‘file3.txt’ and ‘file4.txt’ we need to set failed = TRUE.

batch_run(path, failed = TRUE, ask = FALSE)
batch_report(path)
#> ✔ file.txt [00:00:00.001]
#> ✔ file2.txt [00:00:00.001]
#> ✔ file3.txt [00:00:00.001]
#> ✔ file4.txt [00:00:00.004]
#> Success: 4
#> Failure: 0
#> Remaining: 0
#> 

Clean Up

With all the files successfully processed the only remaining task is to delete the hidden configuration and log files.

list.files(path, all.files = TRUE)
#> [1] "."           ".."          ".batchr.log" ".batchr.rds" "file.txt"   
#> [6] "file2.txt"   "file3.txt"   "file4.txt"
print(batch_cleanup(path))
#>    . 
#> TRUE
list.files(path, all.files = TRUE)
#> [1] "."         ".."        "file.txt"  "file2.txt" "file3.txt" "file4.txt"