Install binary version of old R packages? - r

I have an R script that was written by someone else that I need to run. This script depends on another library (written by yet another person) that generates an error when the script is run. This all began after I updated my desktop to the newest Mac OSX, and subsequently updated to the newest R (and newer versions of these dependencies). I'm unable to compile this older version due to some compilers errors (error: unsupported directive '.stabs'), so a binary version seems like the easier way to go than to start digging deep into OPC as a novice R programmer. Is there a way to force R to install older versions of a binary library like there is for package managers for languages like Python and Ruby?

So, what I ended up doing that worked fine was rsyncing my old installation from an old computer that had those libraries already installed. Both were Mac OSX, but the older computer was still on the previous OS and R was compatible and the scripts do run now.
Also, I did find that if you don't reinstall R after you upgrade OS X, the installation is still valid, you just have to recreate the sym links or add the binary to your PATH.

Related

Unable to load fonts on Ubuntu

Using R version 4.1.2 on Ubuntu 20.04, I run hist(10). The error says
X11 font -adobe-helvetica-%s-%s-*-*-%d-*-*-*-*-*-*-*, face 2 at size 14 could not be loaded
I find two things on google. This thread I can't understand, and this thread suggests I build R differently. I did recently build R from source, without the cairo flag recommended by the Bio-Deago github thread, but I have now replaced it with a binary from Rstudio, and the problem persists. Any suggestions? Thanks!
EDIT: the problem only occurs in Rstudio, not when I run R through a terminal.
R gets installed to different places depending on whether you install from source or as a binary: see this, which says:
For versions of R installed by system package managers (e.g. r-base on Debian or Ubuntu) this will be /usr/bin/R. For versions of R installed from source this will typically (but not always) be /usr/local/bin/R.
Rstudio looks first at which R, which in this case is the version from source. I deleted the source version (target of symlink at which R). Then when I ran R in a terminal, said R is not installed. Then, I followed the Rstudio binary install instructions mentioned in the question. I followed them from the beginning even though I had already installed an Rstudio binary. Now I can run hist(10) and get a plot as expected.

Upgrade R version from 3.4.4 to 3.5.3 on Ubuntu 16.04

When we ssh onto this Linux server, it says Welcome to Ubuntu 16.04.7 LTS (GNU/Linux 4.4.0-193-generic x86_64), so this is the type of server we are on.
When we run R to launch R in the server from terminal, and then version, we get:
What is the easiest way to upgrade from version 3.4.4 to version 3.5.3? We don't want to upgrade to v4 or to the latest version quite yet until we ensure no breaking issues with our codebase on R v4, however we would like to make this upgrade to 3.5.3 so we can run an R library that requires 3.5.3+.
How can we do this from the command line on the server here, and after switching from 3.4.4 to 3.5.3 do we then need to re-install all R libraries on the server as well? I'm worried if I start running stuff from the command line and I mess up that I'll break our current R app.
I would be surprised if there is a .deb available for this OS release/R release combination.
This page only offers R 4.1, but as far back as 16.04 LTS
This page offers older R versions, but only back to 18.04 LTS
It's also possible that one of the backports on the Debian packages page would work on your system (it reports that 3.5.3 is available for Debian "jessie", although I have no idea how that lines up with Ubuntu releases).
I guess it's possible that one of those .debs works for your system though ... ? If you do install from .deb, you may have to be careful not to clobber your current version (e.g. using the --instdir argument to dpkg).
Building from source:
download the source tarball for R 3.5.3
untar, ./configure, make, sudo make install (assuming you have all the necessary development tools, system libraries, etc.; you'll have to troubleshoot/install as you go along)
It might be worth ./configure --prefix=/path/to/testdir to put the new version in a completely separate location, just to make extra-sure you don't stomp on your current install. (This would also simplify the library-path stuff below.)
Packages do need to be re-installed when switching major versions (e.g. 3.4.x to 3.5.x); one way to do it is to copy the system library of packages to a new place (adjust library paths as necessary; see R installation and administration manual).
Then, you should be able to update.packages(checkBuilt=TRUE, ask=FALSE) to update everything.
The tricky spots are (1) getting the library paths right (this will depend a bit on how things are set up in your existing installation; (2) it's conceivable that some current versions of packages on CRAN will fail to re-install/re-build under R 3.5.3. devtools::install_version() would help, although you'd have to track down the correct version manually. I recall someone posting about a package that would install an archived version by date, which could save a lot of poking around ...

How to upgrade RStudio version?

Recently I've tried to install keras package in R. I ran the following commands:
install.packages("keras") library(keras)
But when I ran is_keras_avalable() I got FALSE.
I was told, that the problem with unsuccessful 'keras' installation may come from the old RStudio version, which is installed on my notebook (I work in RStudio 3.4.0). I was adviced to reinstall RStudio and to get a brand new version.
Could you, please, tell me how to do it? I'd be very pleased if you provide me with the necessary links for installing the latest RStudio version on Windows 8.1.
Thank you for your help.
RStudio is an application (specifically, an Integrated Development Environment, or IDE). This is a common tool for using R, which is a computer language. Much like Microsoft Word is a common tool for writing documents in English (or other languages - and in fact you can use RStudio to write files other languages like RMarkdown, Python, SQL, French, or Swahili.). However, just like you can write English in programs other than Word, you can also write and run R code in programs other than RStudio.
You can find the most recent version on the RStudio website, or you can look in the Help menu and select "Check for updates", which will give you a download link if a newer version is available.
Given the version number you show, 3.4.0, you need to upgrade R (the language), not RStudio (the application). The current RStudio version as I write this is 1.3.*, while the current R version is 4.0.*.
The place you need to go to update R is CRAN - their website has download links for Windows, Mac, and Linux.

Can homebrew R and "standard" R for MacOS from CRAN coexist?

I am running R 3.6.1 on a Mac Mini running Sierra and a MacBook Pro running El Capitan. I normally get all the R packages that I need from CRAN or github and use them without issues, but I am trying to install and use an R package (NicheMapR) that requires a fortran compiler and this is giving me issues. Even after installing gfortran, the R package still does not work (the fortran code seems to be compiled but the package installation fails). The package developer suggested that installing R via homebrew might solve the problem. On the contrary, my hunch is that it would lead to a world of pain, to quote Walter from the Big Lebowski. My questions are:
What is the advantage of a homebrew version of R for MacOSX over the "regular" version installed from CRAN?
Can the two versions coexist?
Is the homebrew version going to affect the regular one?
Finally: is homebrew going to help or will it simply open a whole
new can of worms?
Many thanks in advance.
Yes, installing from homebrew is a recipe for pain. It's specifically recommended against by the official CRAN binary maintainer see his remarks from March 2016 on r-sig-mac.
Regarding your questions, this can be summarized as:
What is the advantage of a homebrew version of R for MacOSX over the "regular" version installed from CRAN?
Positives: Select your own BLAS and easily work with geospatial tools.
Downsides: Always needing to compile each R package.
Can the two versions coexist?
Yes. The homebrew version installs into a different directory. But, watch out for library collision (see next question). However, you will have to deal with symbolic linking regarding what version of R is accessible from the console and you will also need to look into using RSwitch to switch between R versions.
Is the homebrew version going to affect the regular one?
Yes, if the library paths overlap. There will be problems regarding package installation and loading. Make sure to setup different library paths. To do so, please look at the .libPaths() documentation.
Finally: is homebrew going to help or will it simply open a whole new can of worms?
Yes and no. Unless you know what you're doing, opt for the CRAN version of R and its assorted goodies.

How to change R versions on Mac

I'm trying to use the legacy version of R to run a package on Mac.
On Windows there is a switch in the General tab. However, it is not found on Mac RStudio. I referred to this post: https://support.rstudio.com/hc/en-us/articles/200486138-Using-Different-Versions-of-R
And I do not quite understand the 3 options:
Run the installer from CRAN for the R version you want to be current
List item Use the RSwitch utility available at: http://r.research.att.com/
Update the R.framework/Versions/Current directory alias directly
using ln -s
Especially for RSwitch utility, how do I even open/use that?
Thanks!
This old question is again timely, with the move to R version 4.
There is an RSwitch.app available for MacOS 10.14+ that provides this functionality via a menu bar app. You can choose which of your R builds to use, integrating with the standard Mac GUI and with RStudio. (I've had trouble with the newest Mac GUI not working with R 3.6.3 but RStudio seems to work fine with both that and R 4.0.1 along with this RSwitch.app.) The help pages for Rswitch show what's going on underneath, if you prefer to work via a command line.
If you use a standard .pkg via GUI to install a newer version of R it will effectively forget the other versions. You can use the command
pkgutil --forget (path to package name)
with the .pkg to overcome that. (I've had trouble installing from a tarball instead; might just be my incompetence.) Installing an older version from a .pkg won't overwrite a newer version but in that case it might be safest to omit any attempt to instal older Tcl/Tk and Texinfo, selectable with the "Customize" button at the "Installation Type" stage of the installation. With security settings in MacOS you might have trouble opening some R .pkg files; the awkward workaround is to use Control/right/two-finger click on the file in the Finder, then select ‘Open With’ and ‘Installer’ from the menu that appears.
The version of RSwitch.app noted in the question is still available via a link from this page. This is evidently 32-bit as it is disabled on my Mac under OS 10.14.

Resources