用于在 Rstudio 中将工作目录设置为源文件位置的 R 命令

Question

I am working out some tutorials in R. Each R code is contained in a specific folder. There are data files and other files in there. I want to open the .r file and source it such that I do not have to change the working directory in Rstudio as shown below:

Is there a way to specify my working directory automatically in R.

Answer 1

To get the location of a script being sourced, you can use utils::getSrcDirectory or utils::getSrcFilename . So changing the working directory to that of the current file can be done with:

setwd(getSrcDirectory()[1])

This does not work in RStudio if you Run the code rather than Source ing it. For that, you need to use rstudioapi::getActiveDocumentContext .

setwd(dirname(rstudioapi::getActiveDocumentContext()$path))

This second solution requires that you are using RStudio as your IDE, of course.

Answer 2

I know this question is outdated, but I was searching for a solution for that as well and Google lists this at the very top:

this.dir <- dirname(parent.frame(2)$ofile)
setwd(this.dir)

put that somewhere into the file (best would be the beginning, though), so that the wd is changed according to that file.

According to the comments, this might not necessarily work on every platform (Windows seems to work, Linux/Mac for some). Keep in mind that this solution is for 'sourcing' the files, not necessarily for running chunks in that file.

see also get filename and path of `source`d file

Answer 3

For rstudio , you can automatically set your working directory to the script directory using rstudioapi like that:

library(rstudioapi)

# Getting the path of your current open file
current_path = rstudioapi::getActiveDocumentContext()$path 
setwd(dirname(current_path ))
print( getwd() )

This works when Running or Sourceing your file.

You need to install the package rstudioapi first. Notice I print the path to be 100% sure I'm at the right place, but this is optional.

Answer 4

dirname(rstudioapi::getActiveDocumentContext()$path)

works for me but if you don't want to use rstudioapi and you are not in a proyect, you can use the symbol ~ in your path. The symbol ~ refers to the default RStudio working directory (at least on Windows).

If your RStudio working directory is "D:/Documents", setwd("~/proyect1") is the same as setwd("D:/Documents/proyect1").

Once you set that, you can navigate to a subdirectory: read.csv("DATA/mydata.csv") . Is the same as read.csv("D:/Documents/proyect1/DATA/mydata.csv") .

If you want to navigate to a parent folder, you can use "../" . For example: read.csv("../olddata/DATA/mydata.csv") which is the same as read.csv("D:/Documents/oldata/DATA/mydata.csv")

This is the best way for me to code scripts, no matter what computer you are using.

Answer 5

I realize that this is an old thread, but I had a similar problem with needing to set the working directory and couldn't get any of the solutions to work for me. Here's what did work, in case anyone else stumbles across this later on:

# SET WORKING DIRECTORY TO CURRENT DIRECTORY:
system("pwd=`pwd`; $pwd 2> dummyfile.txt")
dir <- fread("dummyfile.txt")
n<- colnames(dir)[2]
n2 <- substr(n, 1, nchar(n)-1)
setwd(n2)

It's a bit convoluted, but basically this uses system commands to get the working directory and save it to dummyfile.txt, then R reads that file using data.table::fread. The rest is just cleaning up what got printed to the file so that I'm left with just the directory path.

I needed to run R on a cluster, so there was no way to know what directory I'd end up in (jobs get assigned a number and a compute node). This did the trick for me.

Answer 6

This answer can help:

script.dir <- dirname(sys.frame(1)$ofile)

Note: script must be sourced in order to return correct path

I found it in: https://support.rstudio.com/hc/communities/public/questions/200895567-can-user-obtain-the-path-of-current-Project-s-directory-

The BumbleBee´s answer (with parent.frame instead sys.frame) didn´t work to me, I always get an error.

Answer 7

The solution

dirname(parent.frame(2)$ofile)

not working for me.

I'm using a brute force algorithm, but works:

File <- "filename"
Files <- list.files(path=file.path("~"),recursive=T,include.dirs=T)
Path.file <- names(unlist(sapply(Files,grep,pattern=File))[1])
Dir.wd <- dirname(Path.file)

More easy when searching a directory:

Dirname <- "subdir_name"
Dirs <- list.dirs(path=file.path("~"),recursive=T)
dir_wd <- names(unlist(sapply(Dirs,grep,pattern=Dirname))[1])

Answer 8

If you work on Linux you can try this:

setwd(system("pwd", intern = T) )

It works for me.

Answer 9

The here package provides the here() function, which returns your project root directory based on some heuristics.

Not the perfect solution, since it doesn't find the location of the script, but it suffices for some purposes so I thought I'd put it here.

Answer 10

I understand this is outdated, but I couldn't get the former answers to work very satisfactorily, so I wanted to contribute my method in case any one else encounters the same error mentioned in the comments to BumbleBee's answer.

Mine is based on a simple system command. All you feed the function is the name of your script:

extractRootDir <- function(x) {
    abs <- suppressWarnings(system(paste("find ./ -name",x), wait=T, intern=T, ignore.stderr=T))[1];
    path <- paste("~",substr(abs, 3, length(strsplit(abs,"")[[1]])),sep="");
    ret <- gsub(x, "", path);
    return(ret);
}

setwd(extractRootDir("myScript.R"));

The output from the function would look like "/Users/you/Path/To/Script" . Hope this helps anyone else who may have gotten stuck.

Answer 11

setwd(this.path::here())

works both for sourced and "active" scripts.

Answer 12

I was just looking for a solution to this problem, came to this page. I know its dated but the previous solutions where unsatisfying or didn't work for me. Here is my work around if interested.

filename = "your_file.R"
filepath = file.choose()  # browse and select your_file.R in the window
dir = substr(filepath, 1, nchar(filepath)-nchar(filename))
setwd(dir)

Answer 13

I feel like a mocking bird, but I'm going to say it: I know this post is old, but...

I just recently learned you cannot call the api when running the script from Task Scheduler and a.bat file. I learned this the hard way. Thought those of you who were using any of the methods might like to know that. We run lots of script this way overnight. Just recently changed our path to include the api called so we could "dynamically" set the working directory. Then, when the first one we tried with failed when triggered from Task Scheduler, investigation brought that information about.

This is the actual code that brought this issue to light:

Works beautifully if you're running the script though!

Just adding my two cents as I know people still pull these threads up, thought it might be helpful.

Answer 14

In case you use UTF-8 encoding:

path <- rstudioapi::getActiveDocumentContext()$path
Encoding(path) <- "UTF-8"
setwd(dirname(path))

You need to install the package rstudioapi if you haven't done it yet.

Answer 15

Most GUIs assume that if you are in a directory and "open", double-click, or otherwise attempt to execute an.R file, that the directory in which it resides will be the working directory unless otherwise specified. The Mac GUI provides a method to change that default behavior which is changeable in the Startup panel of Preferences that you set in a running session and become effective at the next "startup". You should be also looking at:

?Startup

The RStudio documentation says:

"When launched through a file association, RStudio automatically sets the working directory to the directory of the opened file." The default setup is for RStudio to be register as a handler for.R files, although there is also mention of ability to set a default "association" with RStudio for.Rdata and.R extensions. Whether having 'handler' status and 'association' status are the same on Linux, I cannot tell.

http://www.rstudio.com/ide/docs/using/workspaces

Answer 16

dirname(parent.frame(2)$ofile)  

doesn't work for me either, but the following (as suggested in https://stackoverflow.com/a/35842176/992088 ) works for me in ubuntu 14.04

dirname(rstudioapi::getActiveDocumentContext()$path)

Answer 17

Here is another way to do it:

set2 <- function(name=NULL) {
  wd <- rstudioapi::getSourceEditorContext()$path
  if (!is.null(name)) {
    if (substr(name, nchar(name) - 1, nchar(name)) != '.R') 
      name <- paste0(name, '.R')
  }
  else {
    name <- stringr::word(wd, -1, sep='/')
  }
  wd <- gsub(wd, pattern=paste0('/', name), replacement = '')
  no_print <- eval(expr=setwd(wd), envir = .GlobalEnv)
}
set2()

Answer 18

I am working out some tutorials in R. Each R code is contained in a specific folder. There are data files and other files in there. I want to open the .r file and source it such that I do not have to change the working directory in Rstudio as shown below:

Is there a way to specify my working directory automatically in R.