roo

Object-oriented programming in R

Object-oriented programming (OOP) can help us encapsulate complex data structures into a simple interface.

OOP concepts

• Encapsulation - related data is stored and handled together
• Polymorphism - the most appropriate function is called based on the object type (e.g various plot functions)
• Inheritance - code reuse by hierarchy of more-to-less general object types

In R, everything has a class (i.e type):

class("Hello")

## [1] "character"

class(1:5)

## [1] "integer"

class(2.1)

## [1] "numeric"

class(table(c(1,4,5,3,2,1)))

## [1] "table"

class(plot)

## [1] "function"

x = 10
class(x)

## [1] "numeric"

OOP in R

Unlike other main-stream languages, R has multiple OOP systems that co-exist in parallel:

• S3 - used in base R
• S4 - de-facto standard in Bioconductor
• Reference classes (RC) - used in special use cases

S3 and S4 are similar, but S4 is more formal. RC is quite different and is Java-like.

Sections

• S3 OOP
• S4 OOP
• RC OOP