R tutorial

R tutorial

This page is a placeholder, or under current development; it is here principally to establish the logical framework of the site. The material on this page is correct, but incomplete.

This is a tutorial introduction to R for users with no previous background in the platform or the language.

 

The environment

In this section we discuss how to download and install the software, how to configure an R session and what work with the R environment includes.

Installation

1. Navigate to http://probability.ca/cran/ and follow the link to your computer's operating system.
2. Download a precompiled binary (or "build") of the R "framework" to your computer and follow the instructions for installing it. You don't need tools, or GUI versions for now, but do make sure that the program is the correct one for your version of your operating system.
3. Launch R.

The program should open a window–the "R console"–and greet you with its input prompt, awaiting your input:

>

The sample code on this page sometimes copies input/output from the console, and sometimes shows the actual commands only. The > character at the beginning of the line is always just R's input prompt; It is shown here only to illustrate the interactive use of the program and you do not need to type it. If a line starts with [1] or similar, this is R's output on the console. A #-character this marks the following text as a comment which is not executed by R. In principle, commands can be copied by you and pasted into the console, or into a script - obviously, you don't need to copy the comments. In addition, I use syntax highlighting on R-script, to color language keywords, numbers, strings, etc. different from other text. This improves readability but keep in mind that the colours you see on your computer will be different. One more thing about the console: use your keyboard's up-arrow keys to retrieve previous commands, then enter the line with left-arrow to edit it; hit enter to execute the modified line.

User interface

R comes with a GUI to lay out common tasks. For example, there are a number of menu items, many of which are similar to other programs you will have worked with ("File", "Edit", "Format", "Window", "Help" ...). All of these tasks can also be accessed through the command line. In general, GUIs are useful when you are not sure what you want to do or how to go about it; the command line is much more powerful when you have more experience and know your way around in principle. R gives you both options.

In addition to the Console, there are a number of other windows that you can open (or that open automatically). They all can be brought to the foreground with the Windows menu and include help, plotting, package browser and other windows.

Let's begin with a glossary of some terms that R uses and how they relate to your work:

Help

Help is available for all commands and for the R command line syntax. As well, help is available to find the names of commands when you are not sure of them.

("help" is a function, arguments to a function are passed in parentheses "()")

> help(rnorm)
>

(shorthand for the same thing)

> ?rnorm
>

(what was the name of that again ... ?)

> ?binom     
No documentation for 'binom' in specified packages and libraries:
you could try '??binom'
> ??binom
>

(found "Binomial" in the list of keywords)

> ?Binomial
>

That's all fine, but you wil soon notice that R's help documentation is not all that helpful for the newcomers (who need the most help). If you look at the bottom of the help function, you will usually find examples of command usage; these often make matters more clear than the terse and principled help-text above. Or you can just Google for what interests you and this is often the quickest way to find working example code. Also, as a result of Google search it may turn out for example that something can't be done (easily)–and you won't find things that can't be done at all in the help system. You may want to include "r language" in your search terms, although Google is usually pretty good at figuring out what kind of "r" you are looking for, if your query includes a few terms vaguely related to statistics.

There is also an active R-help mailing list which you can post to–or at least search the archives: your question probably has been asked and answered before.

Working directory

To locate a file in a computer, one has to specify the filename and the directory in which the file is stored; this is sometimes called the path of the file. The "working directory" for R is either the direcory i which the R-program has been installed, or some other directory, as initialized by a startup script. You can execute the command getwd() to list what the "Working Directory" is currently set to:

> getwd()
[1] "/Users/steipe/R"

It is convenient to put all your R-input and output files into a project specific directory and then define this to be the "Working Directory". Use the setwd() command for this. setwd() requires a parameter in its parentheses: a string with the directory path. Strings in R are delimited with " or ' characters. If the directory does not exist, an Error will be reported. Make sure you have created the directory. On Mac and Unix systems, the usual shorthand notation for relative paths can be used: ~ for the home directory, . for the current directory, .. for the parent of the current directory.

My home directory...

> setwd("~") 
> getwd()
[1] "/Users/steipe"

Relative path: home directory, up one level, then down into chen's home directory)

> setwd("~/../chen")  
> getwd()
[1] "/Users/chen"

Absolute path: specify the entire string)

> setwd("/Users/steipe/abc/R_samples")  
> getwd()
[1] "Users/steipe/abc/R_samples"

Task:

1. Create a directory for your sample files and use setwd("your-directory-name") to set the working directory.
2. Confirm that this has worked by typing getwd().

The Working Directory functions can also be accessed thorugh the Menu, under Misc.

Workspace

During an R session, you might define a large number of variables, datastructures, load packages and scripts etc. All of this information is stored in the so-called "Workspace". When you quit R you have the option to save the Workspace; it will then be reloaded in your next session.

List the current workspace contents: initially it is empty. (R reports an object of type "character" with a length of 0.)

> ls() 
character(0)
>

Initialize three variables (multiple commands on one line can be separated with a semicolon";")

> a <- 1; b <-2; eps <- 0.0001
> ls() 
[1] "a"   "b"   "eps"
>

Remove one item. (Note: the parameter is not the string "a", but the variable name a.)

> rm(a) 
> ls()
[1] "b"   "eps"
>

We can use the output of ls() as input to rm() to remove everything and clear the Workspace. (cf. ?rm for details)

rm(list= ls()) 
> ls() 
character(0)
>

 

Packages

R has many powerful functions built in, but one of it's greatest features is that it is easily extensible. Extensions have been written by legions of scientists for many years, most commonly in the R programming language itself, and made available through CRAN–The Comprehensive R Archive Network. A package is a collection of code, documentation and sample data files. To use packages, you need to install them (once), and add them to your current session (for every new session). You can get an overview of installed and loaded packages by opening the Package Manager window from the Packages & Data Menu item. It gives a list of available packages you currently have installed, and identifies those that have been loaded at startup, or interactively.

Task:

• Navigate to http://cran.r-project.org/web/packages/ and read the page.
• Navigate to http://cran.r-project.org/web/views/ (the CRAN task-views.
• Follow the link to Genetics and read the synopsis of available packages. The library sequinr sounds useful, but check first whether it is already installed.

library() opens a window of installed packages in the library; search() shows which one are currently loaded.

> library()
> search()
 [1] ".GlobalEnv"        "tools:RGUI"        "package:stats"     "package:graphics" 
 [5] "package:grDevices" "package:utils"     "package:datasets"  "package:methods"  
 [9] "Autoloads"         "package:base"

• In the R packages available window, confirm that seqinr is not yet installed.
• Follow the link to seqinr to see what standard information is available with a package. Then follow the link to Reference manual to access the documentation pdf. This is also sometimes referred to as a "vignette" and contains usage hints and sample code.

Read the help for vignette. Note that there is a command to extract R sample code from a vignette, to experiment with it.

> ?vignette
>

Install seqinr from the closest CRAN mirror and load it for this session. Explore some functions.

> ??install
> ?install.packages
> install.packages("seqinr")
--- Please select a CRAN mirror for use in this session ---
trying URL 'http://probability.ca/cran/bin/macosx/contrib/2.13/seqinr_3.0-5.tgz'
Content type 'application/x-gzip' length 4528528 bytes (4.3 Mb)
opened URL
==================================================
downloaded 4.3 Mb


The downloaded packages are in
	/var/folders/dq/dqPEEPbF0ApRU/-Tmp-//RtmpBlw/downloaded_packages
> 
> library("seqinr")
> ls("package:seqinr")
  [1] "a"                       "aaa"                     "AAstat"                 
  [4] "acnucclose"              "acnucopen"               "al2bp"                  
     [...]
[205] "where.is.this.acc"       "words"                   "words.pos"              
[208] "write.fasta"             "zscore"                 
> ?a
> a("Tyr")
[1] "Y"
> choosebank()
 [1] "genbank"       "embl"          "emblwgs"       "swissprot"     "ensembl"      
    [...]
 [31] "refseqViruses"

• The fact that these methods work shows that the library has been downloaded, installed and downloaded and its functions are now available. Just for fun and demonstration, let's use these functions to download a sequence and calculate some statistics (however, not to digress too far, without further explanation at this point). Copy the code below and paste it into the R-console

choosebank("swissprot")
query("seq", "N=MBP1_YEAST")
mbp1 <- getSequence(seq)
closebank()
x <- AAstat(mbp1[[1]])
barplot(sort(x$Compo))

Scripts

My preferred way of running R is not strictly through the console. I open a new file - a script - and enter my R commands into the file. Then I execute the commands directly from the script. I may try things in the console, experiment, change parameters etc. - but ultimately everything I do goes into the file. This has four major advantages:

• The script is an accurate record of my procedure so I know exactly what I have done;
• I add numerous comments to record what I was thinking when I developed it;
• I can immediately reproduce the entire analysis from start and finish, simply by rerunning the script;
• I can reuse parts easily, thus making new analyses quick to develop.

Task:

• Use the File menu to open a New Document.
• Enter the following code (copy from here and paste):

# sample script:
# define a vector
a <- c(1, 1, 2, 3, 5, 8, 13)
# list its contents
a
# calculate the mean of its values
mean(a)

• save the file in your working directory (e.g. with the name sample.R).

Placing the cursor in a line and pressing command-return (on the Mac, ctrl-r on Windows) will execute that line and you see the result on the console. You can also select more than one line and execute the selected block with this shortcut. Alternatively, you can run the entire file. In the console type:

source("sample.R")

However: this will not print output to the console. When you run a script, if you want to see text output you need to explicitly print() it.

• Change your script to the following, save it and source() it.

# sample script:
# define a vector
a <- c(1, 1, 2, 3, 5, 8, 13)
# list its contents
print(a)
# calculate the mean of its values
print(mean(a))

• Confirm that the print(a) command also works when you execute the line directly from the script.

Nb. if you want to save your output to file, you can divert it to a file with the sink() command. You can read about the command by typing:

?sink

 

Simple commands

The R command line evaluates expressions. Expressions can contain constants, variables, operators and functions of the various datatypes R recognizes.

Operators

The common arithmetic operators are recognized in the usual way. Try the following operators on numbers:

5
5 + 3
5 + 1 / 2
3 * 2 + 1
3 * (2 + 1)
2^3 # Exponentiation
8 ^ (1/3) # Third root via exponentiation
7 %% 2  # Modulo operation (remainder of integer division)
7 %/% 2 # Integer division

Functions

Most of R's functionality is expressed through functions. These are either defined by default (built-in), loaded in specific packages (see above), or they can be easily defined by you (see below). In general a function is invoked through a name, followed by one or more arguments (also parameters) in parentheses, separated by commas. Whenever I refer to a function, I write the parentheses to identify it as such and not a constant or other keyword eg. log(). Here are some examples for you to try and play with:

cos(pi) #"pi" is a predefined constant.
sin(pi) # Note the rounding error. This number is not really different from  zero.
sin(30 * pi/180) # Trigonometric functions use radians as their argument - this conversion calculates sin(30 degrees)
exp(1) # "e" is not predefined, but easy to calculate.
log(exp(1)) # functions can be arguments to functions - they are evaluated from the inside out.
log(10000) / log(10) # log() calculates natural logarithms; convert to any base by dividing by the log of the base. Here: log to base 10.
exp(complex(r=0, i=pi)) #Euler's identity

There are several ways to populate the argument list for a function and R makes a reasonable guess what you want to do. Consider the specification of a complex number in Euler's identity above. The function complex() can work with a number of arguments that are given in the documentation (see: ?complex). These include length.out, real, imaginary, and some more. The length.out argument creates a vector with one or more complex numbers. If nothing else is specified, this will be a vector of complex zero(s). If there are two, or three arguments, they will be placed in the respective slots. However, since the arguments are named, we can also define which slot of the argument list they should populate. Consider the following to illustrate this:

complex(1)
complex(4)
complex(1, 2) # imaginary part missing - defaults to zero
complex(1, 2, 3) # one complex number
complex(4, 2, 3) # four complex numbers
complex(real = 0, imaginary = pi) # defining via named parameters
complex(imaginary = pi, real = 0) # same thing - if names are used, order is not important
complex(re = 0, im = pi) # names can be abbreviated ...
complex(r = 0, i = pi) # ... to the shortest string that is unique among the named parameters. Use this with discretion to keep your code readable.
complex(i = pi, 1, 0) # Think: what have I done here? Why does this work?

Variables

In order to store the results of evaluations, you can freely assign them to variables. Variables are created internally whenever you first use them (i.e. they are allocated and instantiated). Variable names are case sensitive. There are a small number of reserved strings, and a very small number of predefined constants, such as pi. However these constants can be overwritten - be careful. Read more at:

?make.names
?reserved

To assign a value to a constant, use the assignment operator <-. You could also use the = sign, but this is too easily confused with the equality test ==, and such errors are hard to debug. Try:

a <- 5
a
a + 3
b <- 8
b
a + b
a == b # equality test
a != b # not equal
a < b  # less than

Note that all of R's data types can be assigned to variables (as well as functions and other objects).

 

Scalar data

Scalars are single numbers, the "atomic" parts of more complex datatypes. We have covered many of the properties of scalars above, e.g. the use of constants and their assignment to variables. To round this off, here are some remarks on the types of scalars R uses and on coercion, or casting one datatype into another. The following scalar types are supported:

• Boolean constants: TRUE and FALSE. This type has the "mode" logical";
• Integers, floats (floating point numbers) and complex numbers. These types have the mode numeric;
• Strings. These have the mode character.

Other modes exist, such as list, function and expression, all of which can be combined into complex objects.

The function mode() returns the mode of an object and typeof() returns its type. Consider:

a <- 3 > 5; a; mode(a); typeof(a) # Note: a > 5 is a logical expression, its value is FALSE.
a <- 3 < 5; a; mode(a); typeof(a)

a <- 3.0;   a;  mode(a); typeof(a) # Double precision floating point number
a <- 3.0e0; a;  mode(a); typeof(a) # Same value, exponential notation

a <- 3;     a;  mode(a); typeof(a) # Note: numbers are double precision floats by default.
a <- as.integer(3);  a;  mode(a); typeof(a) # If we really want an integer, we must coerce to type integer.

a <- "3"; a;  mode(a); typeof(a) # Forcing the number to be interpreted as a character.

# More coercions. For each of these, first think what result you would expect:
as.numeric("3") # character as numeric
as.numeric("3.141592653") # string as numeric
as.numeric("pi") # another string as numeric
as.numeric(pi) # not a string, but a predefined constant

as.logical(0)
as.logical(1)
as.logical(-1)
as.logical(pi) # any non-zero number is TRUE ...
as.logical("pi") # ... but not non-numeric types. NA is "Not Available".

 

Vectors

Since we (almost) never do statistics on scalar quantities, R obviously needs ways to handle collections of data items. In its the simplest form such a collection is a vector: an ordered list of items of the same type. Vectors are created from scratch with the c() function which concatenates individual items into a list. Vectors have properties, such as length; individual items in vectors can be combined in useful ways.

#Create a vector and list its contents and length:
f <- c(1, 1, 3, 5, 8, 13, 21)
f
length(f)

# Various ways to retrieve values from the vector.
f[1] # By index: "1" is first element. 
f[length(f)] # length() is the index of the last element.
1:4 # This is the range operator
f[1:4] # using the range operator (it generates a sequence and returns it in a vector)
f[4:1] # same thing, backwards
seq(from=2, to=6, by=2) # The seq() function is a flexible, generic way to generate sequences
seq(2, 6, 2) # Same thing: arguments in default order
f[seq(2, 6, 2)]

# ...using an index vector with positive indices
a <- c(1, 3, 4, 1) # the elements of index vectors must be valid indices of the target vector. The index vector can be of any length.
f[a] # Here, four elements are retrieved from f[]

# ...using an index vector with negative indices
a <- -(1:4) # If elements of index vectors are negative integers, the corresponding elements are excluded.
f[a] # Here, the first four elements are omitted from f[]
f[-(length(f)-4:length(f))] # Here, the last four elements are omitted from f[]

# ...using a logical vector
f>4 # A logical expression operating on the target vector returns a vector of logical elements. It has the same length as the target vector.
f[f>4]; # We can use this logical vector to extract only elements for which the logical expression evaluates as TRUE

# Example: extending the Fibonacci series for three steps. 
# Think: How does this work? What numbers are we adding here and why does the result end up in the vector?
f <- c(f, f[length(f)-1] + f[length(f)]); f 
f <- c(f, f[length(f)-1] + f[length(f)]); f 
f <- c(f, f[length(f)-1] + f[length(f)]); f

Many operations on scalars can be simply extended to vectors and R computes them very efficiently by iterating over the elements in the vector.

f
f+1
f*2

# computing with two vectors of same length
a <- f[-1]; a # like f[], but omitting the first element
b <- f[1:(length(f)-1)]; b # like f[], but shortened by the least element
c <- a / b # the "golden ratio", phi (~1.61803 or (1+sqrt(5))/2 ), an irrational number, is approximated by the ratio of two consecutive Fibonacci numbers.
c
abs(c - ((1+sqrt(5))/2)) # Calculating the error of the approximation, element by element

 

Matrices

If we need to operate with several vectors, or multi-dimensional data, we make use of matrices or more generally k-dimensional arrays R. Matrix operations are very similar to vector operations, in fact a matrix actually is a vector for which the number of rows and columns have been defined.

The most basic form of such definition is the dim() function. Consider:

a <- 1:12; a
dim(a) <- c(2,6); a
dim(a) <- c(2,2,3); a

Alternatively, matrices can be defined using the matrix() or array() functions (see there), or "glued" together from vectors by rows or columns, using the rbind() or cbind() functions respectively:

a <- 1:4
b <- 5:8
c <- rbind(a, b); c
d <- cbind(a, b); d
e <- cbind(d, 9:12); e

Addressing (retrieving) individual elements or slices from matrices is simply done by specifying the appropriate indices, where a missing index indicates that the entire row or column is to be retrieved

e[1,] # first row
e[,2] # second column
e[3,2] # element at index row=3, column = 2
e[3:4, 1:2] # submatrix

 

Lists

...

 

Data frames

...

 

Data manipulations

• Transformation
• Search
• Sort

 

Writing your own functions

 

Notes

 

Further reading and resources