conda-like env for R

Question

[QX]

Hi all,

Is there any ways to manage multi-projects in R using an environment management like conda for python? I know that conda also allow R packages also, but people said that r packages via conda got a lot of bugs.

[dariober]

people said that r packages via conda got a lot of bugs.

In my experience, I don't recall any particular problem with conda and R packages. Occasionally, when I add a package (R on anything else) to an existing environment I have to remove that environment and build it again to resolve some dependency issues. A bit annoying but no big deal, usually.

I would just go for conda since if you are on this forum chances are that you will need programs other than R that are supported by (bio)conda. Except... forget about conda and use mamba instead.

[zx8754]

Maybe https://rstudio.github.io/renv/articles/renv.html

The renv package helps you create reproducible environments for your R projects.

[dthorbur]

I agree with @dariober, installing conda R packages through conda (e.g., conda install r-dplyr) rather than R's install.package function has always worked reasonably well for me. That said, whole environment solving can be problematic with conda. I would suggest using mamba over conda. It's generally much faster and better at solving environments.