2

I am wishing to take advantage of the language agnostic element of the Conda package manager. However, it does not mirror the CRAN package repository, which is deterring my team from using Conda to manage packages for R.

Is there a work around for this or plans to extend the library? I am interested in installing packages like 'quanteda' and 'xgboost' without installing all the packages independently?

This would allow us to do all our package management in Conda as a team.

EB88
  • 841
  • 1
  • 10
  • 26
  • In general, to install your R dependencies, you specify the `r` channel, e.g., `conda install -c r r-xgboost`. What do you mean "installing all the packages independently"? Each package has a set of dependencies that it brings with it when it is installed, this is handled automatically – darthbith Feb 12 '18 at 19:03
  • Thank you darthbith. xgboost was a bad example as it exists in a conda channel. But what about a library like text2vec. What i mean by "independently" is having to install the dependencies separately. This can be a problem with e.g. python pip packages when there are a lot of external dependencies and consequent package conflicts. Obviously conda addresses this issue, but only if it has the package registered in a repository. – EB88 Feb 13 '18 at 09:01
  • If the package is not available as a conda package, you will have to build your own package. If the package is available on CRAN (or PyPI), getting a simple package is as easy as `conda skeleton cran ` or `conda skeleton pypi `. See: https://conda.io/docs/user-guide/tutorials/build-pkgs-skeleton.html – darthbith Feb 13 '18 at 14:23
  • Duplicate covers essentially same topic with a variety of pertinent answers. It should also be noted that both the example packages here were available on `conda-forge` channel. Lastly, [getting stuff on CRAN added to Conda Forge is very streamlined](https://stackoverflow.com/a/69921221/570918). – merv May 28 '23 at 03:21

2 Answers2

1

What about using conda r channel and packages when available and in addition install r packages from CRAN when they are only available there? This is what our team is doing and has been working fine.

David Tam
  • 485
  • 5
  • 13
0

first download the "R" libraries as tar file from conda site , place into user/anaconda3/"pkgs" folder open conda 3 prompt and type command: conda install "path name /filename"

ravindra mishra
  • 1