R: RGraphviz installation - r

Question

I've been trying to install the library using brute force - trying different combinations of the things people have posted in mailing lists (I'm too lazy to list them out one by one, but I think I tried the most of them. I can list that too if it helps anyone.). The results have varied from a harmless message of a missing dll to RGui not being able to start before I remove the library manually. Nevertheless, I haven't succeeded...
Do you know how to install it properly, so that it works? I'm running 64bit Windows 7 and I'm not keen of compiling packages from source. Thanks!

Answer 1

Install Rgraphviz 2.2.1 from Bioconductor
According to the latest README:
Rgraphviz now comes bundles with Graphviz. This should greatly simplify
installation on all platforms, compared with earlier versions.
Bioconductor 2.11 contains a lot of libraries that you might not want or need, but it does seem to be the easiest path to achieving what you want. These are the instructions on the Rgraphviz homepage:
source("http://bioconductor.org/biocLite.R")
biocLite("Rgraphviz")
These instructions work for R x64 2.15.2 on Windows 7
library("Rgraphviz")
Loading required package: graph
Loading required package: grid
set.seed(123)
V <- letters[1:10]
M <- 1:4
g1 <- randomGraph(V, M, 0.2)
plot(g1)

Answer 2

The README inside the source package for RGraphviz contains very some pretty clear instructions.
I think it's a bit user-unfriendly to those who only want to install the binary package, to hint that they might also want to download and unpack a tar.gz file containing the complete source, in order to find some technical info .... which turns out to be absolutely crucial.

Answer 3

One word of warning. If you follow the installation instructions for Rgrahpviz precisely, things won't work. Installing the package graphviz edits the environment PATH, but incorrectly. I didn't notice and I bet lots of others missed it too.
Rgraphviz wants to look for the binary in ...;C:\Graphviz2.20\bin\;
BUT the graphviz install adds a path only as far as ;C:\Graphviz2.20\;.
You'll have to edit it. Older instructions suggest a manual edit anyway, but newer ones leave it to graphviz.

Answer 4

I have same problems to install Rgraphviz (for using it in Bayesian Network packages). I used to get brute force solutions, but, now I trying an other one described here in this page
maybe you had also tried. If you succeed in installing Rgraphviz, I will be grateful if you learn me how to do.

Answer 5

I am using R-Studio on Ubuntu 12.10 x64 and installed Rgraphviz from the BioC software repository. Hope this helps. Eg: http://www.biotricks.net/2012/03/rstudio-and-bioconductor.html

Answer 6

I don't know if you have managed to solve your problem 100% but my installation problem came with the R version. Rgraphiz is developed by Bioconductor and it seems to be outdated. However, I had to use it since I moved to a new company and they are using it in one of their shiny apps for whatever reason.
With that being said, I have a dirty solve that I came up with after a couple of days of struggling with it.
First I include a line in the code which seems to be needed since without that line the shiny app simply doesn't do anything after opening:
if(!requireNamespace('BiocManager', quietly = TRUE))
install.packages('BiocManager')
It's weird that it won't work otherwise.The above line is found on Bioconductor's site: https://www.bioconductor.org/packages/devel/bioc/html/Rgraphviz.html
The second thing you need to do only once is to run the biocLite command below in the R console to install the Rgraphviz package
source("http://bioconductor.org/biocLite.R")
biocLite("Rgraphviz")
The above command I found on another thread: R: RGraphviz installation
Hopefully this update will help you or anyone else.

Answer 7

Since with the version of R 3.6.1 the script http://bioconductor.org/biocLite.R returns this message " Error: With R version 3.5 or greater, install Bioconductor packages using BiocManager; see https://bioconductor.org/install " I solved the problem with the following steps:
Get the list of directories used by R to install libraries and
choose the one with write permissions using: .libPaths()
Installing the "BiocManager" library using: install.packages("BiocManager")
Installation of the library "bioconductor" by forcing the directory
with write permissions using: BiocManager::install("Rgraphviz", lib = "C:/Users/tizbet/Documents/R/win-library/3.6")
I worked on the following R installation:
platform x86_64-w64-mingw32, arch x86_64, os mingw32, system x86_64, mingw32, version.string R version 3.6.1 (2019-07-05)