Today our existing users use SAS but are moving to R and a Machine Learning environment. They save their SAS datasets (.sas7bdat file) to a Unix file system. We use Samba shares to allow file sharing between Windows and Unix. So each Unix file system has its own Samba share that makes its contents visible to users working on a Windows platform.
Our users want to be able to load SAS dataset files directly into the ML environment using R code. What is the best way to do this / what is the correct package/library to use?
The Samba share will present the underlying Unix file system as any normal Windows UNC path. Our users will need to provide their username and password in the connection to authenticate them via the Samba share. This needs to be done at the point where they are executing the file load/read command inside their R client.
We are already using packages like
rxImport to load SAS files into R.
What we need is either a “single package” that allows us to load SAS dataset files that accepts parameters for the file source, username and password. Or, an expended “multi-stage approach” that first uses some other package/library to establish the connection to the file source (the Samba share) using the appropriate credentials, that is followed up by the loading of the SAS dataset using something like
Thanks and regards … Rob