How do you handle R Data internal to a package?

The R package I am developing requires several R data objects, such as pre-computed models and parameters.

Currently I have each object in the 'data' directory of the package in individual .RData files. When using the package users can use the "data" function to attach these objects to their environment.

The behaviour I would like instead would be that on loading the package the data objects are automatically attached to the internal package environment and not accessible to the user directly.

My understanding is that placing a 'sysdata.rda' file in the 'R' directory of the package containing the objects currently in 'data' will give me the desired result. However, is there a way to do this so that I can have each object in a separate file instead of grouped together?

Answers


You can use the .onLoad() hook to call data() when your package is being loaded, and specify the package namespace as the environment where to load the data objects into.

Assuming you have files model1.R and mydata.RData in the data/ directory of your package called foopkg, define the function

.onLoad <- function(libname, pkgname) {
  data("model1", "mydata", package=pkgname, envir=parent.env(environment()))
}

somewhere in your package (e.g. in foopkg-package.R).

After building and installing the package,

> library(foopkg)
> ls(loadNamespace("foopkg")) 

should demonstrate that the various data objects were successfully loaded into the package namespace, i.e. visible to functions in your package but not polluting the global environment.


Put your sysdata.rda file in the data directory of your package.

Do not use Lazy Data -- your DESCRIPTION file should either not have a line for LazyData, or, if it does, it should be LazyData: no

In any .R file in the R directory of your package add a line like this

data(sysdata, envir=environment())

I created a data.frame named sysdata and saved it to a file called sysdata.rda in the data directory of a package called anRpackage

I added the above line to an .R file, and also added this unexported function just to show that functions in the package have access to the data.

foo <- function() tail(sysdata, 2)

Then I see the following an R session

> library(anRpackage)
> sysdata
Error: object 'sysdata' not found

> anRpackage:::sysdata
  A  B C
1 1  6 a
2 2  7 b
3 3  8 c
4 4  9 d
5 5 10 e

> anRpackage:::foo()
  A  B C
4 4  9 d
5 5 10 e

So, users still have access to the data, but as you requested, they do not have direct access. The user still has the option to run data(sysdata).


Need Your Help

JSLint Error: Unexpected 'this'

javascript constructor jslint

Having trouble understanding why JSLint is surprised by my use of this in the following code:

MessageBodyWriter not found for media type=application/json

java rest jersey jax-rs

I am facing issues while consuming JAX-RS services as JSON.