Optional argument parsing in R for working directory - r

Question

I am parsing argument in a rscript (merge_em.r) below. Let's say I run the code below using commandline Rscript merge_em.r dataframe1, dataframe2 which gives me this error: Error in setwd(working.dir) : character argument expected. I want to keep working directory argument optional. How do I do it?
library("argparse")
merge_em <- function (x, y, working.dir){
mergedfile <- merge (x, y, by = intersect(names(x), names(y)))
if (missing(working.dir)) {
print ("Working directory not specified! Will use present working directory.")
working.dir <- as.character(file.path(getwd()))
} else {
working.dir <- working.dir
}
setwd(working.dir)
write.table (mergedfile, "merged.txt",
col.names = FALSE,
row.names = FALSE,
sep = "\t",
quote = FALSE
)
}
main <- function() {
# breaks if you set warn = 2
options(error = traceback,
warn = 1)
parser <- ArgumentParser(prog = "merge_em.r",
description = "Merge dataframes")
parser <- ArgumentParser()
parser$add_argument("x")
parser$add_argument("y")
parser$add_argument(
"--working_dir",
dest = "working.dir",
type = "character",
metavar = "DIR",
required = FALSE,
help = "Working directory where files are present"
)
args <- parser$parse_args()
working.dir <- args$working.dir
x <- args$x
if (!R.utils::isAbsolutePath(x))
x <- file.path(working.dir, x)
y <- args$y
if (!R.utils::isAbsolutePath(y))
y <- file.path(working.dir, y)
tryCatch(
merge_em (x, y, working.dir)
,
finally = setwd(working.dir)
)
}
main()

Answer 1

You could exchange the missing() conditional to this:
if (working.dir=="") {
print ("Working directory not specified! Will use present working directory.")
working.dir <- as.character(file.path(getwd()))
} else {
print ("Working directory is specified!")
working.dir <- working.dir
}
And change the argument for working_dir to (default=""):
parser$add_argument(
"--working_dir",
dest = "working.dir",
type = "character",
metavar = "DIR",
default="",
required = FALSE,
help = "Working directory where files are present"
)
And change the tryCatch to:
tryCatch(merge_em(x, y, working.dir), finally = print("Fin"))
Why are you using setwd() io the finally part? If the argument is not given, there is nothing to set or?
Like that you can call the script like this, for example:
Rscript merge_em.r data_frame1, data_frame2
Or with a directory:
Rscript merge_em.r data_frame1, data_frame2, --working_dir "path_to_folder"
Full code:
library(argparse)
merge_em <- function (x, y, working.dir){
mergedfile <- merge (x, y, by = intersect(names(x), names(y)))
if (working.dir=="") {
print ("Working directory not specified! Will use present working directory.")
working.dir <- as.character(file.path(getwd()))
} else {
print ("Working directory is specified!")
working.dir <- working.dir
}
setwd(working.dir)
write.csv(x = mergedfile, file = "merged.txt",
row.names = FALSE,
quote = FALSE
)
}
main <- function() {
# breaks if you set warn = 2
options(error = traceback,
warn = 1)
parser <- ArgumentParser(prog = "merge_em.r",
description = "Merge dataframes")
parser <- ArgumentParser()
parser$add_argument("x")
parser$add_argument("y")
parser$add_argument(
"--working_dir",
dest = "working.dir",
type = "character",
metavar = "DIR",
default="",
required = FALSE,
help = "Working directory where files are present"
)
args <- parser$parse_args()
working.dir <- args$working.dir
x <- args$x
if (!R.utils::isAbsolutePath(x))
x <- file.path(working.dir, x)
y <- args$y
if (!R.utils::isAbsolutePath(y))
y <- file.path(working.dir, y)
tryCatch(merge_em(x, y, working.dir), finally = print("Fin"))
}
main()

Answer 2

You can set it as a default and override it when necessary.
merge_em <- function (x, y, working.dir = getwd()){
mergedfile <- merge (x, y, by = intersect(names(x), names(y)))
setwd(working.dir)
write (mergedfile, "merged.txt",
col.names = FALSE,
row.names = FALSE,
sep = "\t",
quote = FALSE
)
}
And override it with some other value:
merger_em(x, y, 'another/path/dir')
I haven't tested this, but default parameters are a standard in many languages.
Also, you can setwd with getwd like: setwd(getwd())