There are quite a few things that one might want to change from the defaults of R. For example, I prefer a certain editor when writing a function, and there are a number of libraries that I will need sooner or later, so I would like to have them loaded. Also, over the years I have written a number of small routines that I use for various tasks, and I want them available at any time.
The two routines to set up things the way I want are
Note that the . in front means that you can’t see this object when you run ls(). You can however with
ls(all.names = TRUE)
The .First is part of the .RData file whereas the .Rprofile is separate.
When R starts it looks for a file called .Rprofile and executes any commands therein. Then it runs the routine .First.
Well, at least that is how it used to be when we were using the command console. For reasons never explained (and much complained about by the users) RStudio ignores the .First file at startup, so we will need to have a work around.
Let’s start with a simple .Rprofile. This is a stand-alone file usually located in your default working directory. You can find out what that is by running
dir("~")
On machines running Win10 it is usually the C:/Users/YourName/Documents folder.
Exactly in what folder the .Rprofile should be in is bit of a mystery, it depends not just on your operating system but even on its version. You might need to do a bit of trial and error!
So here is what (part of) mine looks like:
options(show.signif.stars=FALSE) # for p-values
options(stringsAsFactors=FALSE) # a classic source of errors
library(ggplot2)
library(wolfr) # Two of my own
library(moodler)
.MyEnv <- new.env()
.MyEnv$sc <- function() source(“clipboard”)
.MyEnv$dp <- function(x) { dump(x,“clipboard”) }
.MyEnv$ip <- function(x) { # Install and immediately load a package
install.packages(x, lib = “C:/R/library”)
library(x, character.only = TRUE)
}
attach(.MyEnv)
if(“First.R” %in% dir()) {# check whether there is a .First
source(“First.R”)
.First()
}
cat(“\nSuccessfully loaded .Rprofile at”, date(), “\n”)
So it changes a few options, loads some libraries, makes a new environment and defines a few functions. Finally it checks the directory from where the Rproject was started to see whether it contains a file First.R. If so it loads and executes it as well.
Say this project is about writing a shiny app, then the First.R would have the line
library(shiny)
Notice because a single .Rprofile sits in the default working directory the same one gets executed every time I start RStudio, but different working directories might have different First.R’s (or none)
Disclaimer: I discuss here Dropbox, but there are other companies that offer the same service and this is not meant as an endorsement of Dropbox (although I do like it myself!).
Dropbox is a cloud based storage site. For me it helps to solve a number of issues:
backup: done automatically
version control: Dropbox keeps all old versions, so if (when!!!) you save a file and then see that this was a mistake you can always go back and find a previous (good) version.
keep things consistent: I have a number of computers, I don’t want to have to work on keeping them synchronized. When I change a file in a Dropbox folder on one computer and then go to another, as soon as the new file is uploaded I got it there as well.
availability: I can get to my files anyplace, anytime.
deal with submissions of students, avoids email
so I have a folder named R in the Dropbox folder. Inside that I have many folders, one for each project.
There is a bit of an issue with this: every minute or so Dropbox wants to synchronize, but RStudio “sees” this and shows an annoying pop-up. To avoid that, open Dropbox by right-clicking the icon on the task bar, select the settings icon, choose preferences, click on the Sync button, selective sync, and uncheck the box of the Rproj folder.
Now of course Dropbox won’t save anything automatically, so don’t forget to do it regularly yourself!