Problems with installation

[JimmyTree]

Hi,
I am trying to install the package but encounter the following error message:
Error : package 'gurobi' is not installed for 'arch = i386'
Error: loading failed
Execution halted
*** arch - x64
ERROR: loading failed for 'i386'

• removing 'C:/Users/jimmyt/Documents/R/R-3.4.1/library/bestsubset'
  Installation failed: Command failed (1)

Also, which version of Gurobi does this package work with? Would the package works with the latest version of gurobi as well?
I am using gurobi 7.0.1 and R 3.4.1 and wondering if this would work?
Thank you very much.

[zhan-gao]

Hi @JimmyTree , I met the same problem, and solved the problem by

library(devtools)
options(devtools.install.args = "--no-multiarch")   
install_github(repo="ryantibs/best-subset", subdir="bestsubset")

The installation will not invoke i386 arch with the additional option.

[JosHageman]

Thanks for sharing! it works!

[BartLoonstra]

Hmm.. It still doesn't work for me. I copy-pasted:

install.packages('devtools')
library(devtools)
options(devtools.install.args = "--no-multiarch")
install_github(repo="ryantibs/best-subset", subdir="bestsubset")

But my error was the same as before:

ERROR: dependency 'gurobi' is not available for package 'bestsubset'

• removing 'D:/R/R-3.4.1/library/bestsubset'
  Installation failed: Command failed (1)

[BartLoonstra]

What I found I had to do:

• Install gurobi optimizer and get the appropriate licence:
  http://www.gurobi.com/downloads/download-center
• Install the package that was somehwere found in the gurobi folder after installation:
  install.packages('C:/gurobi751/win64/R/gurobi_7.5-1.zip', repos=NULL)
• Install devtools:
  install.packages('devtools')
  library(devtools)
• Install a package for gurobi called 'slam' (library not needed; as it is called automatically by gurobi):
  install.packages('slam')
• Run the code to install the 'bestsubset' package from github:
  options(devtools.install.args = "--no-multiarch")
  install_github(repo="ryantibs/best-subset", subdir="bestsubset")

Result:
* DONE (bestsubset)

[KiraDynnes]

Hi all

I followed @BartLoonstra recipe above and it worked perfect - until I updated R to the new 3.5.0.
Installation of gurobi works fine install.packages
> install.packages('C:/gurobi800/win64/R/gurobi_8.0-0.zip', repos=NULL)
Installing package into ‘C:/Users/kdsv/Documents/R/win-library/3.5’ (as ‘lib’ is unspecified) package ‘gurobi’ successfully unpacked and MD5 sums checked
but I cannot seem to load the library:
> library(gurobi)
Error: package or namespace load failed for ‘gurobi’: package ‘gurobi’ was installed by an R version with different internals; it needs to be reinstalled for use with this R version
I uninstalled gurobi and installed i again - nu luck, then I downloaded the 8.0 version and tried again - still no luck. Any suggestions will be very much appreciated!

[javierfajnolla]

I think that this is a problem with gurobi's R package, which does not seem to install or work in R 3.5. The one provided with gurobi 8.0 does work with previous versions.

I was in the same situation as you. I have written to gurobi, but it is been a week and I did not get any reply from them. The only solution I can think about until they update the binaries is to go back to the previous R version. If you had it installed, maybe you have not uninstalled it yet and can point RStudio (or whatever you are using) to this old instalation...

[KiraDynnes]

Im back with 3.4.4 - and it all works like a charm! I guess that is what you get for being a little too fresh on your updates ;-)

[javierfajnolla]

Awesome :)

For the record, if someone else ends up here, I was able to install gurobi's R package from source to work with R 3.5 by doing this: (I did this on a mac, but it should be very similar on windows or linux, as the files for those OS are also there)

• Download source code from: http://files.gurobi.com/gurobi800/gurobiR.tar.gz (I got this address from gurobi's assistance mail)

• Unzip the file somewhere.

• Go inside unzipped folder from a terminal (or CMD in windows) and type:
  make -f Makefile.mac (or Makefile.win for windows; or Makefile.linux)

• This creates the zipped package that can be installed within R: gurobi_8.0-0.tgz.

• So, in R or RStudio, just install the this file, either from RStudio option or with the command:
  install.packages("/Library/gurobi800/mac64/R/from_source/gurobiR/gurobi_8.0-0.tgz", repos = NULL)
  (where "/Library/gurobi800/mac64/R/from_source/gurobiR/gurobi_8.0-0.tgz" is the address to the zipped file that we have just created)

Hope it helps!

[vtshen]

I am able to install the packages bestsubset and gurobi following this online guide https://cran.r-project.org/web/packages/prioritizr/vignettes/gurobi_installation.html , but only in R in Linux terminal.

In R studio, when I am running library(bestsubset), the error message is

"unable to load shared object '/home/shen/R/x86_64-pc-linux-gnu-library/3.4/gurobi/libs/gurobi.so': libgurobi80.so: cannot open shared object file: No such file or directory"

The official document (http://www.gurobi.com/documentation/8.0/quickstart_linux.pdf) mentions that

If you are using R from RStudio Server, and you get an error indicating that R is unable to load
the Gurobi DLL or shared object, you may need to set the rsession-ld-library-path entry in the server config file. Please consult the RStudio documentation for more information.

I tried to add

rsession-ld-library-path=/opt/gurobi/linux64/lib

in each of the two created files rserver.conf and rsession.conf in the folder /etc/rstudio, but not working. Any suggestions?

[faengl]

I had the same problem on Ubuntu 16.04.4. with the desktop version of RStudio (not RStudio Server). My problem was solved by adding these lines to ~/.profile as suggested in this post.
Notice that for the variable GUROBI_HOME you have to substitute the correct path on your system.

export GUROBI_HOME="/path/to/installation/dir/gurobi801/linux64"
export PATH="${PATH}:${GUROBI_HOME}/bin"
export LD_LIBRARY_PATH="${LD_LIBRARY_PATH}:${GUROBI_HOME}/lib"

[Pancreatologist]

Hi everyone, i find the way to solve this problem in my computer. Just install the gurobi current version and get the free license for non-commerical use only. and in the R studio.
install.packages('c:/gurobi1000/win64/R/gurobi_10.0-0.zip', repos=NULL)
I hope it might help you.