Dear all,
There are different versions of R. Packages depend to their released time need newer or older version of R.
There are some explanation in the r website as "md5sum of the .exe to the true fingerprint". However, i couldnt get what exactly should be done!.
It would be nice to have your advice about switching from different versions of R and running different packages.
Warm regards,
Mehdi
Dear Hamidreza,
Thanks for your comments. Some codes unfortunately work only with older versions. I am not sure update always be the best choice.
Dear Mehdi!
You can use the older version of a package, you have to install it from the source.
For example:
packageurl
The problem exists when a new version of R does not support an existing Package. Then you cannot install that Package via CRAN and you probably have to install it via GitHub, if you are lucky and a repository exists.
NOt sure if it's the best way, but i have two different R versions installed in my machine, to make sure the package i need are in the right version.
if you want to stay reproducible and take care of using different versions of R and R packages with different projects, you might take snapshots of R. The package packrat might be a good choice for it.
There has been a report on R-Bloggers considering this topic, see http://www.r-bloggers.com/introducing-the-reproducible-r-toolkit-and-the-checkpoint-package/
This might be interesting for your question
update: http://www.r-bloggers.com/updated-checkpoint-package-faster-reproducibility-with-more-feedback/
Hello Mehdi, from tools options in R studio there is option called general from this you can change the version of R. If you want more control use renv package.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package