rjulia provides an interface between R and Julia. It allows a user to run a script in Julia from R, and maps objects between the two languages.
It currently supports use on Linux and Windows (both R and RGUI), but build on Windows only for advance users.
Install julia v0.5 and R version >=3.1.0.
Add <juliainstalldir>/bin
to your system PATH variable if needed.
Install the rjulia package. If you're using the v0.4 branch of julia, use the "julia0.4" branch of rjulia.
You can install rjulia on Linux using the devtools package:
install.packages("devtools") #if not already installed
devtools::install_github("armgong/rjulia", ref="julia0.5")# or ref="julia0.4" if using Julia v0.5
You can install rjulia on Windows using the Rtools and devtools package :
install.packages("devtools") #if not already installed
devtools::install_github("armgong/rjulia", ref="master", args = "--no-multiarch")# or ref="0.3" if using Julia v0.3
If you want to be able to use R or Julia objects that contain NA values or factors or data frames, the Julia packages DataArrays
and DataFrames
must be installed.
library(rjulia)
julia_init() #**(will auto find your julia home)**
julia_eval("1+1")
Please see the *.R
files in the demo/
directory, or use
demo(package = "rjulia")
Help files are now done, mostly with examples.
The Julia api rapidly changes between releases. Each time you upgrade or downgrade Julia, rjulia needs to be recompiled and reinstalled, e.g. with devtools::install_github
.
Due to RStudio issue (https://github.com/armgong/rjulia/issues/16), when using rjulia on Windows 64bit, RStudio-0.98.1103 is recommended http://download1.rstudio.org/RStudio-0.98.1103.zip .
Due to gcc toolchain issue , when using rjulia on Windows, it will crash R, to solve this, download unoffical build julia from https://github.com/armgong/julia-64-build-with-mingw-builds-for-R .