[R] Is it good practice to have a package depend on Rtools(for unziping of .7z and.rar files)? Crab legal?

Lucas Ferreira Mation lucasmation at gmail.com
Thu Mar 29 15:56:52 CEST 2018

I created this package,


which we are preparing for Cran submission. The package facilitates
downloading and reading most of the microdatasets (household surveys,
Census, etc). For each dataset it:

(a) downloads the data from  the data providers
(b) "unzip" the necessary files
(c) imports the data into R.

For  (b), some of the datasets come in .7z and .rar format which ideally
the package should unzip (or whatever the correct them for uncompacting
these file types is) so that the underlying .txt or .csv files can be
imported into R.

Not being able to find a solution within R (R only deals with .zip files),
we added a Rtools dependency, in order to use Rtools expanded
unzipping capabilities on the .7z and .rar  files.

Is adding a Rtools dependency a bad practice for an R package? Is it even
allowed by cran? I mean, it changes requires the installation of a
standalone software (Rtools), which does not seem very user-friendly.

Is there any alternative for unzipping .rar and .7z files that only depends
on R packages and works independently of the OS, etc?

If the Rtools is a big problem, I could also revert to our initial strategy
that did not support .7z and .rar "unzipping", requesting the user to
manually unpack those files.


	[[alternative HTML version deleted]]

More information about the R-help mailing list