Building R packages Sren Hjsgaard Department of Mathematical - - PowerPoint PPT Presentation

Building R packages

Søren Højsgaard Department of Mathematical Sciences Aalborg University, Denmark September 24, 2012

Printed: September 24, 2012 File: package-slides.tex

2

Contents

1 Why build R packages 3 2 Getting started 4 3 EXERCISES: Writing R packages 5

3

1 Why build R packages

To have R–work made publically available on CRAN http://cran.r-project.org/ are, the work must be put into an R package. Another reason for creating an R package is that it is a simple way of organizing your own work – and of sharing work among collaborators. Windows users must download and install Rtools from http://cran.r-project.org/. For linux users, no particular steps are necessary. For Mac users, I do not know.

4

2 Getting started

A skeleton of a package called ASC can be created with

R> package.skeleton("ASC")

The ASC directory contains an R directory (where R functions are put) and a man directory (where documentation files are put).

5

To install the package, do R CMD INSTALL ASC To build the package (create a tar–ball which can be distributed), do R CMD build ASC To create a zip file (the easiest way to install a package on windows) do R CMD INSTALL --build ASC To check that the package meets the requirements of CRAN to R CMD check ASC Then there are many additional bells and whistles that can be added. See“Writing R Extensions”for details: http://cran.r-project.org/doc/manuals/R-exts.html

6

3 EXERCISES: Writing R packages

• 1. Create a template for a package called ASC. Make sure that you can compile

and install the package.

• 2. Put the R functions you have created so far in the course into the package.

Create documentation files for your functions. (Hint: prompt() is your friend). Create documentation files for your datasets. (Hint: promptData() is your friend).