Paradigms  Multiparadigm: procedural, objectoriented, functional, reflective, imperative, array^{[1]} 

Designed by  Ross Ihaka and Robert Gentleman 
Developer  R Core Team 
First appeared  August 1993 
Stable release  4.1.0^{[2]}
/ 18 May 2021 
Typing discipline  Dynamic 
License  GNU GPL v2 
Filename extensions 

Website  www 
Influenced by  
 
Influenced  
Julia^{[6]}  

R is a programming language and free software environment for statistical computing and graphics supported by the R Core Team and the R Foundation for Statistical Computing.^{[7]} The R language is widely used among statisticians and data miners for developing statistical software and data analysis. Polls, data mining surveys, and studies of scholarly literature databases show substantial increases in R's popularity;^{[8]} as of July 2021,^{[update]} R ranks 12th in the TIOBE index, a measure of popularity of programming languages.^{[9]}
The official R software environment is a GNU package. It is written primarily in C, Fortran, and R itself (thus, it is partially selfhosting) and is freely available under the GNU General Public License. Precompiled executables are provided for various operating systems. Although R has a command line interface, there are several thirdparty graphical user interfaces, such as RStudio, an integrated development environment, and Jupyter, a notebook interface.
R is an implementation of the S programming language combined with lexical scoping semantics, inspired by Scheme.^{[1]} S was created by John Chambers in 1976 while at Bell Labs. A commercial version of S was offered as SPLUS starting in 1988.
Much of the code written for SPLUS runs unaltered in R.^{[10]}
In 1991 Ross Ihaka and Robert Gentleman at the University of Auckland, New Zealand, began an alternative implementation of the basic S language, completely independent of SPLUS, which they began publicizing in 1993.^{[11]} It was named partly after the first names of the first two R authors and partly as a play on the name of S.^{[12]} In 1995, Martin Maechler convinced Ihaka and Gentleman to make R free and opensource software under Version 2 of the GNU General Public License.^{[13]}^{[14]} The R Core Team was formed in 1997 to further develop the language.^{[12]} As of 2021^{[update]}, it consisted of Gentleman, Ihaka, and Maechler, plus Douglas Bates, John Chambers, Peter Dalgaard, Kurt Hornik, Tomas Kalibera, Michael Lawrence, Friedrich Leisch, Uwe Ligges, Thomas Lumley, Martin Morgan, Paul Murrell, Martyn Plummer, Brian Ripley, Deepayan Sarkar, Duncan Temple Lang, Luke Tierney, and Simon Urbanek. Heiner Schwarte, Guido Masarotto, Stefano Iacus, Seth Falcon, and Duncan Murdoch were also formerly members.^{[15]}
The first official release came in 1995.^{[11]} The Comprehensive R Archive Network (CRAN) was officially announced 23 April 1997 with 3 mirrors and 12 contributed packages.^{[16]} The first official "stable beta" version (v1.0) was released 29 February 2000.^{[17]}^{[18]}
R and its libraries implement various statistical and graphical techniques, including linear and nonlinear modeling, classical statistical tests, spatial and timeseries analysis, classification, clustering, and others. R is easily extensible through functions and extensions, and the R community is noted for its active contributions in terms of packages. Many of R's standard functions are written in R itself,^{[citation needed]} which makes it easy for users to follow the algorithmic choices made. For computationally intensive tasks, C, C++, and Fortran code can be linked and called at run time. Advanced users can write C, C++,^{[19]} Java,^{[20]} .NET^{[21]} or Python code to manipulate R objects directly.^{[22]} R is highly extensible through the use of usersubmitted packages for specific functions or specific areas of study. Due to its S heritage, R has stronger objectoriented programming facilities than most statistical computing languages.^{[citation needed]} Extending R is also eased by its lexical scoping rules.^{[23]}
Another strength of R is static graphics, which can produce publicationquality graphs, including mathematical symbols. Dynamic and interactive graphics are available through additional packages.^{[24]}
R has Rd, its own LaTeXlike documentation format, which is used to supply comprehensive documentation, both online in a number of formats and in hard copy.^{[25]}
R is an interpreted language; users typically access it through a commandline interpreter. If a user types 2+2
at the R command prompt and presses enter, the computer replies with 4, as shown below:
> 2 + 2
[1] 4
This calculation is interpreted as the sum of two singleelement vectors, resulting in a singleelement vector. The prefix [1]
indicates that the list of elements following it on the same line starts with the first element of the vector (a feature that is useful when the output extends over multiple lines).
Like other similar languages such as APL and MATLAB, R supports matrix arithmetic. R's data structures include vectors, matrices, arrays, data frames (similar to tables in a relational database) and lists.^{[26]} Arrays are stored in columnmajor order.^{[27]} R's extensible object system includes objects for (among others): regression models, timeseries and geospatial coordinates. The scalar data type was never a data structure of R.^{[28]} Instead, a scalar is represented as a vector with length one.^{[29]}
Many features of R derive from Scheme. R uses Sexpressions to represent both data and code.^{[citation needed]} Functions are firstclass and can be manipulated in the same way as data objects, facilitating metaprogramming, and allow multiple dispatch. Variables in R are lexically scoped and dynamically typed.^{[30]}</ref> Function arguments are passed by value, and are lazy—that is to say, they are only evaluated when they are used, not when the function is called.^{[31]}
R supports procedural programming with functions and, for some functions, objectoriented programming with generic functions.^{[32]} A generic function acts differently depending on the classes of arguments passed to it. In other words, the generic function dispatches the function (method) specific to that class of object. For example, R has a generic print
function that can print almost every class of object in R with a simple print(objectname)
syntax.^{[33]}
Although used mainly by statisticians and other practitioners requiring an environment for statistical computation and software development, R can also operate as a general matrix calculation toolbox – with performance benchmarks comparable to GNU Octave or MATLAB.^{[34]}
Main article: R package 
The capabilities of R are extended through usercreated^{[35]} packages, which allow specialised statistical techniques, graphical devices, import/export capabilities, reporting tools (RMarkdown, knitr, Sweave), etc. The large number of packages available for R, and the ease of installing and using them, has been cited as a major factor in driving the widespread adoption of the language in data science.^{[36]}^{[37]}^{[38]}^{[39]}^{[40]} The R packaging system is also used by researchers to create compendia to organise research data, code and report files in a systematic way for sharing and public archiving.^{[41]}
A core set of packages is included with the installation of R, with more than 15,000 additional packages (as of September 2018^{[update]}) available at the Comprehensive R Archive Network (CRAN),^{[42]} Bioconductor, Omegahat,^{[43]} GitHub, and other repositories.^{[44]}
The "Task Views" page (subject list) on the CRAN website^{[45]} lists a wide range of tasks (in fields such as Finance, Genetics, High Performance Computing, Machine Learning, Medical Imaging, Social Sciences and Spatial Statistics) to which R has been applied and for which packages are available. R has also been identified by the FDA as suitable for interpreting data from clinical research.^{[46]}
Other R package resources include RForge,^{[47]} a central platform for the collaborative development of R packages, Rrelated software, and projects. RForge also hosts many unpublished beta packages, and development versions of CRAN packages. Microsoft maintains a daily snapshot of CRAN, that dates back to Sept. 17, 2014.^{[48]}
The Bioconductor project provides R packages for the analysis of genomic data. This includes objectoriented datahandling and analysis tools for data from Affymetrix, cDNA microarray, and nextgeneration highthroughput sequencing methods.^{[49]}
A group of packages called the Tidyverse, which can be considered a "dialect of the R language", is increasingly popular in the R ecosystem.^{[note 1]} The group of packages strives to provide a cohesive collection of functions to deal with common data science tasks, including data import, cleaning, transformation and visualisation (notably with the ggplot2 package).
R is one of 5 languages with an Apache Spark API, the others being Scala, Java, Python, and SQL.^{[50]}^{[51]}
A list of changes in R releases is maintained in various "news" files at CRAN.^{[52]} Some highlights are listed below for several major releases.
Release  Date  Description 

0.16  This is the last alpha version developed primarily by Ihaka and Gentleman. Much of the basic functionality from the "White Book" (see S history) was implemented. The mailing lists commenced on April 1, 1997.  
0.49  19970423  This is the oldest source release which is currently available on CRAN.^{[53]} CRAN is started on this date, with 3 mirrors that initially hosted 12 packages.^{[54]} Alpha versions of R for Microsoft Windows and the classic Mac OS are made available shortly after this version.^{[citation needed]} 
0.60  19971205  R becomes an official part of the GNU Project. The code is hosted and maintained on CVS. 
0.65.1  19991007  First versions of update.packages and install.packages functions for downloading and installing packages from CRAN.^{[55]} 
1.0  20000229  Considered by its developers stable enough for production use.^{[56]} 
1.4  20011219  S4 methods are introduced and the first version for Mac OS X is made available soon after. 
1.8  20031008  Introduced a flexible condition handling mechanism for signalling and handling condition objects. 
2.0  20041004  Introduced lazy loading, which enables fast loading of data with minimal expense of system memory. 
2.1  20050418  Support for UTF8 encoding, and the beginnings of internationalization and localization for different languages. 
2.6.2  20080208  Last version to support Windows 95, 98, Me and NT 4.0^{[57]} 
2.11  20100422  Support for Windows 64bit systems. 
2.12.2  20110225  Last version to support Windows 2000^{[58]} 
2.13  20110414  Adding a new compiler function that allows speeding up functions by converting them to bytecode. 
2.14  20111031  Added mandatory namespaces for packages. Added a new parallel package. 
2.15  20120330  New load balancing functions. Improved serialisation speed for long vectors. 
3.0.0  20130403  Support for numeric index values 2^{31} and larger on 64bit systems. 
3.3.3  20170306  Last version to support Microsoft Windows XP. 
3.4.0  20170421  Justintime compilation (JIT) of functions and loops to bytecode enabled by default. 
3.5.0  20180423  Packages bytecompiled on installation by default. Compact internal representation of integer sequences. Added a new serialisation format to support compact internal representations. 
3.6.0  20190426  Improved sampling from a discrete uniform distribution, which was noticeably nonuniform on large populations.^{[59]} New serialisation format supported since 3.5.0 becomes the default. 
4.0.0  20200424  R now uses a stringsAsFactors = FALSE default, and hence by default no longer converts strings to factors in calls to data.frame() and read.table() . Reference counting is used for tracking object sharing, which reduces the need for copying objects. New syntax for raw string constants.

4.1.0  20210518  Introduced > as the pipe operator for base R syntax (similar to the %>% operator of the magrittr package) and the anonymous function shortcut syntax \(x) x+1

There are various applications than can be used to edit or run R code.^{[60]}
While historically users prefer to run R via the command line console^{[61]} it can also be run using an IDE. IDEs for R include (in alphabetical order) Rattle GUI, R Commander, RKWard, RStudio, and TinnR.^{[61]} R is also supported in multipurpose IDEs such as Eclipse via the StatET plugin,^{[62]} and Visual Studio via the R Tools for Visual Studio.^{[63]}
Editors that support R include Emacs, Vim (NvimR plugin),^{[64]} Kate,^{[65]} LyX,^{[66]} Notepad++,^{[67]} Visual Studio Code, WinEdt,^{[68]} and TinnR.^{[69]} Jupyter Notebook can also be configured to edit and run R code.^{[70]}
R functionality is accessible from several scripting languages such as Python,^{[71]} Perl,^{[72]} Ruby,^{[73]} F#,^{[74]} and Julia.^{[75]} Interfaces to other, highlevel programming languages, like Java^{[76]} and .NET C#^{[77]}^{[78]} are available as well.
The main R implementation is written in R, C, and Fortran,^{[79]} and there are several other implementations aimed at improving speed or increasing extensibility. A closely related implementation is pqR (pretty quick R) by Radford M. Neal with improved memory management and support for automatic multithreading. Renjin and FastR are Java implementations of R for use in a Java Virtual Machine. CXXR, rho, and Riposte^{[80]} are implementations of R in C++. Renjin, Riposte, and pqR attempt to improve performance by using multiple processor cores and some form of deferred evaluation.^{[81]} Most of these alternative implementations are experimental and incomplete, with relatively few users, compared to the main implementation maintained by the R Development Core Team.
TIBCO built a runtime engine called TERR, which is part of Spotfire.^{[82]}
Microsoft R Open is a fully compatible R distribution with modifications for multithreaded computations.^{[83]}
R has local communities worldwide for users to network, share ideas, and learn.^{[84]}^{[85]}
There are a growing number of R events bringing its users together, such as conferences (e.g. useR!, WhyR?, conectaR, SatRdays),^{[86]}^{[87]} meetups,^{[88]} as well as RLadies groups^{[89]} that promote gender diversity and the R Foundation taskforce on women and other underrepresented groups.^{[90]}
The official annual gathering of R users is called "useR!".^{[91]} The first such event was useR! 2004 in May 2004, Vienna, Austria.^{[92]} After skipping 2005, the useR! conference has been held annually, usually alternating between locations in Europe and North America.^{[93]} Subsequent conferences have included:^{[91]}
Future conferences planned are as follows:^{[91]}^{[94]}
The R Journal is the open access, refereed journal of the R project for statistical computing. It features short to medium length articles on the use and development of R, including packages, programming tips, CRAN news, and foundation news.
R is comparable to popular commercial statistical packages such as SAS, SPSS, and Stata, but R is available to users at no charge under a free software license.^{[95]}
In January 2009, the New York Times ran an article charting the growth of R, the reasons for its popularity among data scientists and the threat it poses to commercial statistical packages such as SAS.^{[96]} In June 2017 data scientist Robert Muenchen published a more indepth comparison between R and other software packages, "The Popularity of Data Science Software".^{[97]}
R is more proceduralcode oriented than either SAS or SPSS, both of which make heavy use of preprogrammed procedures (called "procs") that are builtin to the language environment and customized by parameters of each call. R generally processes data inmemory, which limits its usefulness in processing extremely large files.^{[98]}
Although R is an opensource project supported by the community developing it, some companies strive to provide commercial support and extensions for their customers. This section gives some examples of such companies.
In 2007, Richard Schultz, Martin Schultz, Steve Weston and Kirk Mettler founded Revolution Analytics to provide commercial support for Revolution R, their distribution of R, which also includes components developed by the company. Major additional components include: ParallelR, the R Productivity Environment IDE, RevoScaleR (for big data analysis), RevoDeployR, web services framework, and the ability for reading and writing data in the SAS file format.^{[99]} Revolution Analytics also offer a distribution of R designed to comply with established IQ/OQ/PQ criteria which enables clients in the pharmaceutical sector to validate their installation of REvolution R.^{[100]} In 2015, Microsoft Corporation completed the acquisition of Revolution Analytics.^{[101]} and has since integrated the R programming language into SQL Server 2016, SQL Server 2017, SQL Server 2019, Power BI, Azure SQL Managed Instance, Azure Cortana Intelligence, Microsoft ML Server and Visual Studio 2017.^{[102]}
In October 2011, Oracle announced the Big Data Appliance, which integrates R, Apache Hadoop, Oracle Linux, and a NoSQL database with Exadata hardware.^{[103]} As of 2012^{[update]}, Oracle R Enterprise^{[104]} became one of two components of the "Oracle Advanced Analytics Option"^{[105]} (alongside Oracle Data Mining).^{[citation needed]}
IBM offers support for inHadoop execution of R,^{[106]} and provides a programming model for massively parallel indatabase analytics in R.^{[107]}
Tibco offers a runtimeversion R as a part of Spotfire.^{[108]}
Mango Solutions offers a validation package for R, ValidR,^{[109]}^{[110]} to make it compliant with drug approval agencies, like FDA. These agencies allow for the use of any statistical software in submissions, if only the software is validated, either by the vendor or sponsor itself.^{[111]}
The following examples illustrate the basic syntax of the language and use of the commandline interface. (An expanded list of standard language features can be found in the R manual, "An Introduction to R".^{[112]})
In R, the generally preferred assignment operator is an arrow made from two characters <
, although =
can be used in some cases.^{[113]}^{[114]}
> x < 1:6 # Create a numeric vector in the current environment
> y < x^2 # Create vector based on the values in x.
> print(y) # Print the vector’s contents.
[1] 1 4 9 16 25 36
> z < x + y # Create a new vector that is the sum of x and y
> z # return the contents of z to the current environment.
[1] 2 6 12 20 30 42
> z_matrix < matrix(z, nrow=3) # Create a new matrix that turns the vector z into a 3x2 matrix object
> z_matrix
[,1] [,2]
[1,] 2 20
[2,] 6 30
[3,] 12 42
> 2*t(z_matrix)2 # Transpose the matrix, multiply every element by 2, subtract 2 from each element in the matrix, and return the results to the terminal.
[,1] [,2] [,3]
[1,] 2 10 22
[2,] 38 58 82
> new_df < data.frame(t(z_matrix), row.names=c('A','B')) # Create a new data.frame object that contains the data from a transposed z_matrix, with row names 'A' and 'B'
> names(new_df) < c('X','Y','Z') # set the column names of new_df as X, Y, and Z.
> print(new_df) #print the current results.
X Y Z
A 2 6 12
B 20 30 42
> new_df$Z #output the Z column
[1] 12 42
> new_df$Z==new_df['Z'] && new_df[3]==new_df$Z # the data.frame column Z can be accessed using $Z, ['Z'], or [3] syntax, and the values are the same.
[1] TRUE
> attributes(new_df) #print attributes information about the new_df object
$names
[1] "X" "Y" "Z"
$row.names
[1] "A" "B"
$class
[1] "data.frame"
> attributes(new_df)$row.names < c('one','two') ## access and then change the row.names attribute; can also be done using rownames()
> new_df
X Y Z
one 2 6 12
two 20 30 42
One of R's strengths is the ease of creating new functions. Objects in the function body remain local to the function, and any data type may be returned.^{[115]} Here is an example usercreated function:
# Declare function “f” with parameters “x”, “y“
# that returns a linear combination of x and y.
f < function(x, y) {
z < 3 * x + 4 * y
return(z) ## the return() function is optional here
}
> f(1, 2)
[1] 11
> f(c(1,2,3), c(5,3,4))
[1] 23 18 25
> f(1:3, 4)
[1] 19 22 25
The R language has builtin support for data modeling and graphics. The following example shows how R can easily generate and plot a linear model with residuals.
> x < 1:6 # Create x and y values
> y < x^2
> model < lm(y ~ x) # Linear regression model y = A + B * x.
> summary(model) # Display an indepth summary of the model.
Call:
lm(formula = y ~ x)
Residuals:
1 2 3 4 5 6
3.3333 0.6667 2.6667 2.6667 0.6667 3.3333
Coefficients:
Estimate Std. Error t value Pr(>t)
(Intercept) 9.3333 2.8441 3.282 0.030453 *
x 7.0000 0.7303 9.585 0.000662 ***

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
Residual standard error: 3.055 on 4 degrees of freedom
Multiple Rsquared: 0.9583, Adjusted Rsquared: 0.9478
Fstatistic: 91.88 on 1 and 4 DF, pvalue: 0.000662
> par(mfrow = c(2, 2)) # Create a 2 by 2 layout for figures.
> plot(model) # Output diagnostic plots of the model.
Short R code calculating Mandelbrot set through the first 20 iterations of equation z = z^{2} + c plotted for different complex constants c. This example demonstrates:
C
, Z
and X
.install.packages("caTools") # install external package
library(caTools) # external package providing write.gif function
jet.colors < colorRampPalette(c("red", "blue", "#007FFF", "cyan", "#7FFF7F",
"yellow", "#FF7F00", "red", "#7F0000"))
dx < 1500 # define width
dy < 1400 # define height
C < complex(real = rep(seq(2.2, 1.0, length.out = dx), each = dy),
imag = rep(seq(1.2, 1.2, length.out = dy), dx))
C < matrix(C, dy, dx) # reshape as square matrix of complex numbers
Z < 0 # initialize Z to zero
X < array(0, c(dy, dx, 20)) # initialize output 3D array
for (k in 1:20) { # loop with 20 iterations
Z < Z^2 + C # the central difference equation
X[, , k] < exp(abs(Z)) # capture results
}
write.gif(X, "Mandelbrot.gif", col = jet.colors, delay = 100)