Basic data types

This chapter introduces the basic data types that are used in R. We mainly show how to create data of these types. There is much more on how to manipulate data in the following chapters.

The most important basic (or “primitive”) data types are the “numeric” (for numbers) and “character” (for text) types. Additional types are the “integer”, which can be used to represent whole numbers; the “logical” for TRUE/FALSE, and the “factor” for categorical variables. These are all discussed below. In later chapters you will see how these basic types can be combined to represent more complex data types.

Numeric and integer values

Let’s create a variable a that is a vector of one number.

a <- 7

To do this yourself, type the code in an R console. Or, if you use R-Studio, use ‘File / New File / R script’ and type it in the new script. Then press “Run” or “Ctrl-Enter” (Apple-Enter on a Mac) to run the line (make sure your cursor is on the line that you want to run).

The “arrow” <- was used to assign the value 7 to variable a. You can pronounce the above as “a becomes 7”.

It is also possible to use the = sign.

a = 7

Although you can use =, we use <- because the arrow clearly indicates the assignment action, and because = is also used in another context (to pass arguments to functions).

The variable name a used above is entirely arbitrary. We could have used x, varib, fruit or any other name that would help us recognize it. There are a few restrictions: variable names cannot start with a number, and they cannot contain spaces or “special” characters, such as * (which is used for multiplication).

To check the value of a, we can ask R to show or print it.

show(a)
## [1] 7
print(a)
## [1] 7

This is also what happens if you simply type the variable name.

a
## [1] 7

In R, all basic values are stored as a vector, a one-dimensional array of n values of a certain type. Even a single number is a vector (of length 1). That is why R shows that the value of a is [1] 7. Because 7 is the first element in vector a.

We can use the class function to find out what type of object a is (what class it belongs to).

class(a)
## [1] "numeric"

numeric means that a is a vector of real (decimal) numbers. Its value is equivalent to 7.000, but trailing zeros are not printed by default. In a few cases it can be useful, or even necessary, to use integer (whole number) values. To create a vector with a single integer you can either use the as.integer function, or append an L to the number.

a <- as.integer(7)
class(a)
## [1] "integer"
a <- 7L
class(a)
## [1] "integer"

There are several ways to create vectors that consists of multiple numbers. For example, you can use the c (combine) function and spell out the values:

b <- c(1.25, 2.9, 3.0)
b
## [1] 1.25 2.90 3.00

But if you want to create a regular sequence of whole numbers, it is easier to use :.

d <- 5:9
d
## [1] 5 6 7 8 9

You can also use the : to create a sequence in descending order.

6:2
## [1] 6 5 4 3 2

The seq function provides more flexibility. For example it allows for step sizes different than one. In this case we go from 3 to 12, taking steps of 3. Try some variations!

seq(2,5,1)
## [1] 2 3 4 5
seq(from=6, to=12, by=3)
## [1]  6  9 12

To go in descending order the by argument needs to be negative.

seq(from=12, to=0, by=-4)
## [1] 12  8  4  0

You can also reverse the order of a sequence, after making the sequence, by using the rev function.

s <- seq(from=0, to=12, by=4)
s
## [1]  0  4  8 12
r <- rev(s)
r
## [1] 12  8  4  0

We will discuss functions like seq in more detail later. But, in essence, a function is a named procedure that performs a certain task. In the examples above, a function name is seq, and the task is to create a sequence of numbers.

The exact specification of the sequence is modified by the arguments that are provided to seq, in this case: from, to, and by. If you are unsure what a function does, or which arguments are available, then read the function’s help page. You can get to the help page for seq by typing ?seq or help(seq), and likewise for all other functions in R.

The rep (for repeat) function provides another way to create a vector of numbers. You can repeat a single number, or a sequence of numbers.

rep(9, times=5)
## [1] 9 9 9 9 9
rep(5:7, times=3)
## [1] 5 6 7 5 6 7 5 6 7
rep(5:7, each=3)
## [1] 5 5 5 6 6 6 7 7 7

Character values

A character variable is used to represent letters, codes, or words. Character values are often referred to as a “string”.

x <- "Yi"
y <- "Wong"
class(x)
## [1] "character"
x
## [1] "Yi"

To distinguish a character value from a variable name, it needs to be quoted. "x" is a character value, but x is a variable! Double-quoted "Yi" is the same as single-quoted 'Yi', but you cannot mix the two in one value: "Yi' is not valid. You can enclose one type of quote inside a pair of the other type. For example, you can do "Yi's dog" or 'Wong said "good bye" and left'.

One of the most common mistakes for beginners is to forget the quotes.

Yi
## Error in eval(expr, envir, enclos): object 'Yi' not found

The error occurs because R tries to print the value of variable Yi, but there is no such variable. So remember that any time you get the error message object 'something' not found, the most likely reason is that you forgot to quote a character value. If not, it probably means that you have misspelled, or not yet created, the variable that you are referring to.

Keep in mind that R is case-sensitive: a is not the same as A. In most computing contexts, a and A are entirely different and, for most intents and purposes, unrelated symbols.

Now let’s create variable countries holding a character vector of five elements.

countries <- c("China", "China", "Japan", "South Korea", "Japan")
class(countries)
## [1] "character"
countries
## [1] "China"       "China"       "Japan"       "South Korea" "Japan"

The function length tells us how long the vector is (how many elements it has).

length(countries)
## [1] 5

If you want to know the number of characters of each element of the vector, you can use nchar.

nchar(countries)
## [1]  5  5  5 11  5

nchar returns a vector of integers with the same length as x (5). Each number is the number of characters of the corresponding element of countries. This is an example of why we say that most functions in R are vectorized. This means that you normally do not need to tell R to compute things for each individual element in a vector.

It is handy to know that letters (a constant value, like pi) returns the alphabet (LETTERS returns them in uppercase), and toupper and tolower can be used to change case.

z <- letters
z
##  [1] "a" "b" "c" "d" "e" "f" "g" "h" "i" "j" "k" "l" "m" "n" "o" "p" "q" "r" "s"
## [20] "t" "u" "v" "w" "x" "y" "z"
up <- toupper(z)
up
##  [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" "K" "L" "M" "N" "O" "P" "Q" "R" "S"
## [20] "T" "U" "V" "W" "X" "Y" "Z"

Perhaps the most commonly used function for string manipulation is paste. This function is used to concatenate strings. For example:

girl <- "Mary"
boy <- "John"
paste(girl, "talks to", boy)
## [1] "Mary talks to John"

By default, paste uses a space to separate the elements. You can change that with the sep argument.

paste(girl, "likes", boy, sep = " ~ ")
## [1] "Mary ~ likes ~ John"

Sometimes you do not want any separator. You can then use sep="" or the paste0 function.

By using the “collapse” argument, we can concatenate all values of a vector into a single value.

paste(countries, collapse=" -- ")
## [1] "China -- China -- Japan -- South Korea -- Japan"

We’ll leave more advanced manipulation of strings for later, but here are two more important functions. To get a part of a string use substr.

substr("Hello World", 1, 5)
## [1] "Hello"
substr("Hello World", 7, 11)
## [1] "World"

To replace characters in a string use gsub or sub.

gsub("l", "!!", "Hello World")
## [1] "He!!!!o Wor!!d"
gsub("Hello", "Bye bye", "Hello World")
## [1] "Bye bye World"

To find elements that fit a particular pattern use grep or grepl. grep returns the index of the matching elements in a vector. You can use the index to subset the original vector (we will see more of this later).

d <- c("az20", "az21", "az22", "ba30", "ba31", "ab32")
i <- grep("ba", d)
i
## [1] 4 5
d[i]
## [1] "ba30" "ba31"
# or like this
grep("ba", d, value=TRUE)
## [1] "ba30" "ba31"

Above, also note the use of #. Lines that start with this character are ignored by R; so they can be used to provide natural language comments.

# Instead of the index, get logical values with grepl
i <- grepl("ba", d)
i
## [1] FALSE FALSE FALSE  TRUE  TRUE FALSE
# return the cases of d for which i is TRUE
d[i]
## [1] "ba30" "ba31"

Which elements of d include the character “2”?

grep("2", d)
## [1] 1 2 3 6

Which elements of d end with the character “2”? $ has a special meaning.

grep("2$", d)
## [1] 3 6

Which elements of d start with the character “b”? ^ has a special meaning.

grep("^b", d)
## [1] 4 5

Logical values

A logical (or Boolean) value is either TRUE or FALSE. They are used very frequently in R and in computer programming in general.

z <- FALSE
z
## [1] FALSE
class(z)
## [1] "logical"
z <- c(TRUE, TRUE, FALSE)
z
## [1]  TRUE  TRUE FALSE

TRUE and FALSE can be abbreviated to T and F, but that is bad practice. This is because it is possible to change the value of T and F to something else — and that would be extraordinarily confusing. In contrast, TRUE and FALSE are constants that cannot be changed.

Logical values are often the result of a computation. For example, here we ask if the values of x are larger than 3, which is TRUE for values 4 and 5

x <- 2:5
x > 3
## [1] FALSE FALSE  TRUE  TRUE

Likewise we can test for equality using two equal signs == (not a single = which would be an assignment!). <= means “smaller or equal” and >= means “larger or equal”.

x == 3
## [1] FALSE  TRUE FALSE FALSE
x <= 2
## [1]  TRUE FALSE FALSE FALSE

Logical values can be treated as numerical values. TRUE is equivalent to 1 and FALSE to 0.

y <- TRUE
y + 1
## [1] 2

However, if you go the other way, only zero is equivalent to FALSE while any number that is not zero, is TRUE

as.logical(0)
## [1] FALSE
as.logical(1)
## [1] TRUE
as.logical(2.5)
## [1] TRUE

Factors

A factor is a nominal (categorical) variable with a set of known possible values called levels. They can be created using the as.factor function. In R you typically need to convert (cast) a character variable to a factor to identify groups for use in statistical tests and models.

f1 <- as.factor(countries)
f1
## [1] China       China       Japan       South Korea Japan
## Levels: China Japan South Korea

But numbers can also be used. For example, they may simply indicate group membership.

f2 <- c(5:7, 5:7, 5:7)
f2
## [1] 5 6 7 5 6 7 5 6 7
f2 <- as.factor(f2)
f2
## [1] 5 6 7 5 6 7 5 6 7
## Levels: 5 6 7

Dealing with factors can be tricky. For example f2 created above is not what it may seem. We see numbers 5, 6 and 7, but these are now just labels to identify groups. They cannot be used in algebraic expressions.

We can convert factors to something else. Here we use as.integer. If you want a number with decimal places, you can use as.numeric instead.

f2
## [1] 5 6 7 5 6 7 5 6 7
## Levels: 5 6 7
as.integer(f2)
## [1] 1 2 3 1 2 3 1 2 3

The result of as.integer(f2) may have been surprising. But it should not be, as there is no direct link between a category with label "5" and the number 5. In this case, "5" is simply the label of the first category and hence it gets converted to the integer 1. Nevertheless, we can get the numbers back as there is an established link between the character symbol "5" and the number 5. So we first create characters from the factor values, and then numbers from the characters.

fc2 <- as.character(f2)
fc2
## [1] "5" "6" "7" "5" "6" "7" "5" "6" "7"
as.integer(fc2)
## [1] 5 6 7 5 6 7 5 6 7

This is different from as.integer(f2), which returned the indices of the factor values. It has no way of knowing if you want factor level 6 to represent the number 6.

At this point it is OK if you are confused about factors and why you might do such things as conversion from and to them.

Missing values

All basic data types can have “missing values”. These are represented by the symbol NA for “Not Available”. For example, we can have vector ‘m’

m <- c(2, NA, 5, 2, NA, 2)
m
## [1]  2 NA  5  2 NA  2

Note that NA is not quoted (it is a special symbol, it is not the word “NA”).

Properly treating missing values is very important. The first question to ask when they appear is whether they should be missing (or did you make a mistake in the data manipulation?). If they should be missing, the second question becomes how to treat them. Can they be ignored? Should the records with NAs be removed?

Time

Representing time is a somewhat complex problem. There are different calendars, hours, days, months, and leap years to consider. As a basic introduction, here is simple way to create date values.

d1 <- as.Date("2015-4-11")
d2 <- as.Date("2015-3-11")
class(d1)
## [1] "Date"
d1 - d2
## Time difference of 31 days

And there are more advanced classes as well that capture date and time.

as.POSIXlt(d1)
## [1] "2015-04-11 UTC"
as.POSIXct(d1)
## [1] "2015-04-11 UTC"

See http://www.stat.berkeley.edu/~s133/dates.html for more info.