The installed directory is not writable, cannot update the packages 'boot', 'class', 'KernSmooth', 'mgcv', 'nnet', 'rpart', 'spaces'

Question

I installed Bioconductor packages on R version 3.1.2 on Ubuntu 14.04 and received the following message:

Downloaded source packages are located in '/ Tmp / RtmpZYw0Qp / loaded_packages Warning message: the directory is installed is not writable, cannot update the packages "boot", "class", "KernSmooth", "mgcv", "nnet", "rpart", "spatial"

What does this mean and how does it affect the use of the above packages on R?

Answer 1

I encountered the same problem as in R-studio. The solution is to give root access to R.

On Windows, this means that you need to run the program as an administrator.

On Unix / Linux, you must run R from the terminal by running sudo R

Then, as soon as you work with sufficient privileges, you can try installing your package, and it should work.

Answer 2

A warning means that more recent versions of the packages are available, but your version cannot be updated because the directory in which the packages are installed cannot be written. This usually means that R was installed with "system administrator" privileges, but you are trying to update packages as a regular user.

The consequence of this is that functions or bug fixes implemented in updated packages are not available to you. Specific effects can range from minor typographical corrections to the manual page through serious bug fixes. Qualitatively, my bet is that for most purposes this does not matter.

The solution is to upgrade these packages when R starts as a user with permission to write to packages in installed directories - usually the last output element .libPaths() or finer with installed.packages()[, "LibPath"]

Answer 3

Today I faced the same problem when installing three Bioconductor packages under Windows. Two (dependencies of the package that I really wanted) were already on my system in newer forms (as described in @ Martin-Morgan's solution), and therefore there was no action. However, one of them was not installed. For this third package, I had success installing from a local file that was downloaded during an unsuccessful installation without the need to escalate privileges (which was my next step, as described in @Ninadmw).

In R, go to the Packages/Install Package(s) from local files menu and navigate to the specified local download directory, which in your case was /tmp/RtmpZYw0Qp/downloaded_packages , and select the package you want to install.

In RStudio (which you should use) go to the Tools\Install Packages menu, change the installation from the field to Package Archive File (.zip; .tar.gz) , use Browse... to go to the specified download directory, which is in your The case was /tmp/RtmpZYw0Qp/downloaded_packages and select the package you want to install. Then click the Install button.