10 Data structures beyond the vector

Data in R may be stored in a multitude of object types, but the most important ones are vector, matrix, list, data frame and tibble.

10.1 Matrices

A matrix is a two-dimensional object that displays data of the same type (numeric, character, etc) in the form of a table. It is built up with the function matrix() in which the data is imported either in the form of combined data elements (ex: c(12, 54, 987, 5, ...)), a series or sequence of data elements (ex: 1:25), or a vector (ex: temperature). In addition, one must define the number of rows and columns with nrow = and ncol =.

In the following example, the object neo is a matrix made of 4 rows and 6 columns filled with the numeric values stored in the vector temperature that we have previously created.

temperature <- c(
  8.7, 9.2, 9.4, 9.5, 9.7, 10.1, 10.3, 10.6, 10.7, 10.8, 11.3, 11.9, 12.2, 12.3,
  11.7, 10.2, 10.3, 10.3, 10.4, 10.3, 10.1, 9.7, 9.5, 9.4)
neo <- matrix(temperature, nrow = 4, ncol = 6)
neo

     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]  8.7  9.7 10.7 12.2 10.3 10.1
[2,]  9.2 10.1 10.8 12.3 10.3  9.7
[3,]  9.4 10.3 11.3 11.7 10.4  9.5
[4,]  9.5 10.6 11.9 10.2 10.3  9.4

10.1.1 Accessing data elements

In a matrix, each row is numbered [x, ] and each column is numbered [ , y]. Any of the data elements may be retrieved by using its coordinates [x, y] preceded by the name of the matrix:

neo[2, 3]

[1] 10.8

A full row or column may be retrieved with the same expression, but we leave empty the coordinate that is not needed:

neo[2, ]

[1]  9.2 10.1 10.8 12.3 10.3  9.7

neo[, 3]

[1] 10.7 10.8 11.3 11.9

10.1.2 About the use of matrices

The use of matrices on this website is very limited. However, you may meet matrices in other projects, so it is best to know about their existence. You can read more about matrices here.

Exercise

Make the matix neo as above, then

select the first two columns
the third column. What happened?
the second row.
The element in the third column second row.

10.2 Arrays

Matrices are two dimensional objects; arrays generalise this to have any number of dimensions. Arrays are an efficient way to store and manipulate high dimensional data (e.g,. latitude * longitude * time), but can be difficult to understand and don’t play well with the tidyververse.

array(data = 1:24, dim = c(2, 4, 3))

, , 1

     [,1] [,2] [,3] [,4]
[1,]    1    3    5    7
[2,]    2    4    6    8

, , 2

     [,1] [,2] [,3] [,4]
[1,]    9   11   13   15
[2,]   10   12   14   16

, , 3

     [,1] [,2] [,3] [,4]
[1,]   17   19   21   23
[2,]   18   20   22   24

10.3 Lists

A list is an object that contains values of one or several data types. It can not only contain single data elements, but also other objects such as vectors, matrices, etc.

Lists are created by the function list() that combines objects. list() conveniently allows for naming the elements by the mean of the symbol =.

In the example below, we will store 6 elements and name them string, number, temp, boolean, words and matrix. Among these elements to be stored are vec.char, temperature and neo, 3 objects that we have created further above on this page.

my_list <- list(
  string = "one",
  number = 2,
  temp = temperature,
  boolean = TRUE,
  words = c("dog", "cat", "fish"),
  matrix = neo
)
my_list

$string
[1] "one"

$number
[1] 2

$temp
 [1]  8.7  9.2  9.4  9.5  9.7 10.1 10.3 10.6 10.7 10.8 11.3 11.9 12.2 12.3 11.7
[16] 10.2 10.3 10.3 10.4 10.3 10.1  9.7  9.5  9.4

$boolean
[1] TRUE

$words
[1] "dog"  "cat"  "fish"

$matrix
     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]  8.7  9.7 10.7 12.2 10.3 10.1
[2,]  9.2 10.1 10.8 12.3 10.3  9.7
[3,]  9.4 10.3 11.3 11.7 10.4  9.5
[4,]  9.5 10.6 11.9 10.2 10.3  9.4

10.3.1 Retrieving list elements

You can access list items by position with [[ notation.

my_list[[1]] # get first element

[1] "one"

Naming elements is quite convenient as it allows you to retrieve them rapidly by the mean of the symbol $. The syntax is as follows: list_name$element_name.
Here we retrieve the element matrix in the list my_list:

my_list$matrix

     [,1] [,2] [,3] [,4] [,5] [,6]
[1,]  8.7  9.7 10.7 12.2 10.3 10.1
[2,]  9.2 10.1 10.8 12.3 10.3  9.7
[3,]  9.4 10.3 11.3 11.7 10.4  9.5
[4,]  9.5 10.6 11.9 10.2 10.3  9.4

10.3.2 Retrieving single data elements

Even better, you can retrieve a single data element contained in a list element. Here you will have to write an expression that makes use of both the symbol $ and the brackets [ ]in the proper order. The syntax is as follows: list_name$element_name[data].
In this first example, we retrieve the data element located at the third position of the object named words in the list my_list:

my_list$words[3]

[1] "fish"

In the second example, we retrieve the data element located at the second row and third column of the matrix named matrix in the list my_list:

my_list$matrix[2, 3]

[1] 10.8

You may read more information about lists here.

Exercise

Make a list with three named elements of different types.

use square bracket notation to extract the second element
use $ notation to extract an element by name.

10.4 Data frames and tibbles

A data frame is a two-dimensional object that stores data of various types in the form of a table. Data frames are a popular way to store research data since the columns are usually associated with single variables (and are thus of a specific data type such as numeric or character) and rows are associated with observations.

Until recently, data frames were the main storage objects for research data. Nowadays, tibbles (an evolution of the data frame that appeared in the tidyverse) are replacing data frames as they are more practical for handling data sets (you will understand why further below). Because of this trend, we will focus mainly on tibbles here in this section and further on this website. It is however likely that you will meet data frames in the course of your studies. Do not worry as we will see how to transform data frames into tibbles.

Tibbles are standard introduced in tidyverse, so you must make sure that the package is active before using these objects. Simply run this command first:

library(tidyverse)

── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.4     ✔ readr     2.1.5
✔ forcats   1.0.0     ✔ stringr   1.5.1
✔ ggplot2   3.5.1     ✔ tibble    3.2.1
✔ lubridate 1.9.3     ✔ tidyr     1.3.1
✔ purrr     1.0.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors

If you have not installed the package yet, have a look at the Section 5.6.2.

10.4.1 Data frames vs. tibbles

The object df printed below is a data frame that stores the average temperature recorded monthly in 2017, 2018 and 2019 at Lygra (Vestland, Norway). It is created with the function data.frame().

df <- data.frame(
  Year = rep(2017:2019, each = 12),
  Month = rep(month.name, 3),
  Avg_temperature = c(
    3.4, 2.8, 4.2, 5.8, 11.4, 12.6, 14.6, 13.9, 13.7, 9.2, 4.3, 3.1, 2.3, 0.5,
    0.8, 6.7, 13.5, 13.6, 16.2, 13.8, 11.6, 8.0, 6.6, 3.9, 1.7, 4.6, 4.0, 9.1,
    8.8, 13.2, 15.4, 15.8, 11.6, 7.8, 3.6, 4.8)
)
df

   Year     Month Avg_temperature
1  2017   January             3.4
2  2017  February             2.8
3  2017     March             4.2
4  2017     April             5.8
5  2017       May            11.4
6  2017      June            12.6
7  2017      July            14.6
8  2017    August            13.9
9  2017 September            13.7
10 2017   October             9.2
11 2017  November             4.3
12 2017  December             3.1
13 2018   January             2.3
14 2018  February             0.5
15 2018     March             0.8
16 2018     April             6.7
17 2018       May            13.5
18 2018      June            13.6
19 2018      July            16.2
20 2018    August            13.8
21 2018 September            11.6
22 2018   October             8.0
23 2018  November             6.6
24 2018  December             3.9
25 2019   January             1.7
26 2019  February             4.6
27 2019     March             4.0
28 2019     April             9.1
29 2019       May             8.8
30 2019      June            13.2
31 2019      July            15.4
32 2019    August            15.8
33 2019 September            11.6
34 2019   October             7.8
35 2019  November             3.6
36 2019  December             4.8

As you may see, you get at once the whole data set with all 36 rows, the 3 variables, the header with column names and the first column that gives a number to each row.

The object tbl below is a tibble that contains exactly the same observations and variables as df. It is built up by the function tibble().

tbl <- tibble(
  Year = rep(2017:2019, each = 12),
  Month = rep(month.name, 3),
  Avg_temperature = c(
    3.4, 2.8, 4.2, 5.8, 11.4, 12.6, 14.6, 13.9, 13.7, 9.2, 4.3, 3.1, 2.3, 0.5,
    0.8, 6.7, 13.5, 13.6, 16.2, 13.8, 11.6, 8.0, 6.6, 3.9, 1.7, 4.6, 4.0, 9.1,
    8.8, 13.2, 15.4, 15.8, 11.6, 7.8, 3.6, 4.8
  )
)
tbl

# A tibble: 36 × 3
    Year Month     Avg_temperature
   <int> <chr>               <dbl>
 1  2017 January               3.4
 2  2017 February              2.8
 3  2017 March                 4.2
 4  2017 April                 5.8
 5  2017 May                  11.4
 6  2017 June                 12.6
 7  2017 July                 14.6
 8  2017 August               13.9
 9  2017 September            13.7
10  2017 October               9.2
# ℹ 26 more rows

Here, you get a more convenient display of the same data:

only the first 10 rows and the header are displayed,
the number of rows not printed is displayed in the present window (# ... with 26 more rows),
the dimensions of the tibble appear clearly in the header (# A tibble: 36 x 3),
the column names come along with a quick description of the data type (<int> for integer, <chr> for character, <dbl> for double, etc).

All in all, tibbles print much better and give more information than data frames do! They also have more predictable behaviour when extracting data from them.

10.4.2 Retrieving data elements

Similarly to vectors, matrices and lists, one can extract single elements from data frames and tibbles. Here, we use brackets [ ] to do so:

df[3, "Avg_temperature"]

[1] 4.2

tbl[3, "Avg_temperature"]

# A tibble: 1 × 1
  Avg_temperature
            <dbl>
1             4.2

One can also retrieve rows or columns:

df[3, ]

  Year Month Avg_temperature
3 2017 March             4.2

tbl[3, ]

# A tibble: 1 × 3
   Year Month Avg_temperature
  <int> <chr>           <dbl>
1  2017 March             4.2

df[, "Avg_temperature"]

 [1]  3.4  2.8  4.2  5.8 11.4 12.6 14.6 13.9 13.7  9.2  4.3  3.1  2.3  0.5  0.8
[16]  6.7 13.5 13.6 16.2 13.8 11.6  8.0  6.6  3.9  1.7  4.6  4.0  9.1  8.8 13.2
[31] 15.4 15.8 11.6  7.8  3.6  4.8

tbl[, "Avg_temperature"]

# A tibble: 36 × 1
   Avg_temperature
             <dbl>
 1             3.4
 2             2.8
 3             4.2
 4             5.8
 5            11.4
 6            12.6
 7            14.6
 8            13.9
 9            13.7
10             9.2
# ℹ 26 more rows

It is also possible to use the symbol $ to retrieve the content of specific variables:

df$Avg_temperature

 [1]  3.4  2.8  4.2  5.8 11.4 12.6 14.6 13.9 13.7  9.2  4.3  3.1  2.3  0.5  0.8
[16]  6.7 13.5 13.6 16.2 13.8 11.6  8.0  6.6  3.9  1.7  4.6  4.0  9.1  8.8 13.2
[31] 15.4 15.8 11.6  7.8  3.6  4.8

tbl$Avg_temperature

 [1]  3.4  2.8  4.2  5.8 11.4 12.6 14.6 13.9 13.7  9.2  4.3  3.1  2.3  0.5  0.8
[16]  6.7 13.5 13.6 16.2 13.8 11.6  8.0  6.6  3.9  1.7  4.6  4.0  9.1  8.8 13.2
[31] 15.4 15.8 11.6  7.8  3.6  4.8

In Chapter 14 you will see an alternative way of manipulating tibbles with the dplyr package.

10.4.3 Transforming a data frame into a tibble

If you have been previously working with data frames, have been given a data frame to work with, or have imported data using functions that create data frames, you may convert them into tibbles by using as_tibble(). Here we convert the data frame df into a tibble:

df_as_tibble <- as_tibble(df)
df_as_tibble

# A tibble: 36 × 3
    Year Month     Avg_temperature
   <int> <chr>               <dbl>
 1  2017 January               3.4
 2  2017 February              2.8
 3  2017 March                 4.2
 4  2017 April                 5.8
 5  2017 May                  11.4
 6  2017 June                 12.6
 7  2017 July                 14.6
 8  2017 August               13.9
 9  2017 September            13.7
10  2017 October               9.2
# ℹ 26 more rows

You may read more about tibbles here.
You may read more about data frames here.

Exercise

Make a tibble with the first column the months of the year, and the second is a sequence of numbers from 1 to 12.

extract the month column using $ notation

Contributors

Jonathan Soulé
Aud Halbritter
Richard Telford