I have contributed several import functions to R package, which batch read in CSVs from specifically structured directories (such as this).
To give users an example, I'd like to ship such a directory structure of CSVs with the package.
Since these are import functions, it obviously makes no sense to include the data as ready-made *.rda
.
So I did this:
inst/
.keyneson.rd
in man/
as if it were a normal *.rda
dataset.R CMD check
passes without notes/warnings.
Is this ok? Will CRAN accept this? Is there a better way?
Jennifer Bryan suggested on twitter that I should use /inst/extdata
(as @Dason suggested) and place an Rdata
into /data/
.
She did the same for her package, which made it past CRAN.
That's what I did, too, works fine. No complaints from R CMD check
.