RCall.jl can simply be installed with
RCall.jl will automatically install R for you using Conda if it doesn't detect that you have R 3.4.0 or later installed already.
Customizing the R installation
Before installing its own copy of R, the RCall build script (run by
Pkg.add) will check for an existing R installation by looking in the following locations, in order.
R_HOMEenvironment variable, if set, should be the location of the R home directory.
- Otherwise, it runs the
R HOMEcommand, assuming
Ris located in your
- Otherwise, on Windows, it looks in the Windows registry.
- Otherwise, it installs the
To change which R installation is used for RCall, set the
R_HOME environment variable and run
Pkg.build("RCall"). Once this is configured, RCall remembers the location of R in future updates, so you don't need to set
You can set
R_HOME to the empty string
"" to force
Pkg.build to re-run the
R HOME command, e.g. if you change your PATH:
ENV["R_HOME"]="" ENV["PATH"]="....directory of R executable..." Pkg.build("RCall")
R HOME doesn't return a valid R library or
R_HOME is set to
"*", RCall will use its own Conda installation of R.
Should you experience problems with any of these methods, please open an issue.
If you want to install R yourself, rather than relying on the automatic Conda installation, you can use one of the following options:
The current Windows binary from CRAN.
The CRAN .pkg or the homebrew/science tap.
Most Linux distributions allow installation of R from their package manager, however these are often out of date, and may not work with RCall.jl. We recommend that you use the updated repositories from CRAN.
The following will update R on recent versions of Ubuntu:
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-keys E084DAB9 sudo add-apt-repository -y "deb http://cran.rstudio.com/bin/linux/ubuntu $(lsb_release -s -c)/" sudo apt-get update -y sudo apt-get install -y r-base r-base-dev
If you have installed R by some other method (e.g. building from scratch, or files copied but not installed in the usual manner), which often happens on cluster installations, then you may need to set
R_HOME or your
PATH as described above before running
Pkg.build("RCall") in order for the build script to find your R installation. RCall requries R to be installed with its shared library. It could be done with the flag
--enable-R-shlib, consult your server administrator if see if it was the caase.
For some environments, you might also need to specify
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:`R RHOME`/lib"
If you have updated your R installation, you may need to re-run
Pkg.build("RCall") as described above, possibly changing the
R_HOME environment variable first.