Skip to main content

Loading Data into R

One way to use R is by typing your data directly into R this is called "hard coding". A more dynamic approach is to load data from an external data source such as .csv files or databases.

Importing Text Files: CSV & TXT​

CSV (Comma-Separated Values) is a plain-text file that contains data. As the name reveals, CSV files use a comma to separate values. Some files might be named ".csv" without actually using a comma and on the other hand ".txt" files may use a comma as a separator. At the end these files are all plain text and there is no validation if a file adheres to a "csv" standard.

example dataset patient_data.csv
Name, Sex, Age, Blood.type
Joe J., m, 34, A+
Laura M., f, 56, AB+
Isobel P., f, 23, 0+
Nina W., f, 18, B-

The first row of the file contains the column names.

tip

With dir(), you can list existing files in your working directory.

utils & read.csv()​

To import a CSV file, you can use the read.csv() function. It is included with the utils package, which is loaded by default when you start R, so you don't need to install anything.

# Import patient_data.csv: patients
patients <- read.csv("patient_data.csv")

# Print the structure of patients
str(patients)

output:
'data.frame':	4 obs. of  4 variables:
 $ Name      : chr  "Joe J." "Laura M." "Isobel P." "Nina W."
 $ Sex       : chr  "m" "f" "f" "f"
 $ Age       : int  34 56 23 18
 $ Blood.type: chr  "A+" "AB+" "0+" "B-"

read.delim()​

TXT files are plain text documents with no predetermined formatting. To import TXT files, use the read.delim() function. By default, the sep argument is set to "\t" (fields are separated by tabs) and the header argument to TRUE (the first row contains the field names).

In the following example, we will import medication.txt, which contains 3 columns, but the column names are not defined in the first line. That means, we will have to set the argument header to FALSE. The file also uses tabs as field separators.

example dataset pain_medication.txt
Ibuprofen	p.o.	400mg
Paracetamol	p.o.	500mg
Aspirin	p.o.	100mg
Tramadol	p.o.	50mg
Tilidine/Naloxone	p.o.	50/4mg
# import pain_medication.txt
medication <- read.delim("pain_medication.txt", header = FALSE)

# summarize medication
summary(medication)

output:
      V1                 V2                 V3           
 Length:5           Length:5           Length:5          
 Class :character   Class :character   Class :character  
 Mode  :character   Mode  :character   Mode  :character

read.table()​

read.csv() and read.delim() simply invoke read.table() with modified default parameters. If you need to read a file which does not conform with these default parameters you can instead use read.table() directly. Of course, you could also use read.table() for a regular csv file as long as you pass the correct parameters.

tip

With the argument col.names = c() within the read.table() function, you can name the different columns.

# Path to the pain_medication.txt file: path
path <- file.path("data", "pain_medication.txt")

# Import the pain_medication.txt file: medication
medication <- read.table(path, sep="\t", col.names = c("drug", "application", "dose"))

# Call head() on medication
head(medication)

output:
               drug application   dose
1         Ibuprofen        p.o.  400mg
2       Paracetamol        p.o.  500mg
3           Aspirin        p.o.  100mg
4          Tramadol        p.o.   50mg
5 Tilidine/Naloxone        p.o. 50/4mg

Importing Data from Excel Spreadsheets​

There is a multitude of packages for xlsx import, here we will choose one at random. To import data from Excel, you can use the readxl package first. You can simply do this by entering install.packages("readxl") into the R console. Then you can load the readxl package by using library(readxl):

install.packages("readxl")
library(readxl)

Before you start importing data from Excel, it can be useful to first print out the names of the sheets in the Excel document. For this, you can use the function excel_sheets().

With the function read_excel(), you can actually import data into R.

note
  • By default, the first sheet of the Excel document will be imported.
  • With read_excel("excelname.xlsx", sheet = 2), you can import the second sheet.
  • Alternatively, you could use read_excel("excelname.xlsx", sheet = "nameofsheet").

Example:

example dataset patient_data.xlsx, sheet 2
NameSexAgeBlood type
Hannah P.f25B+
Josh T.m33B+
Emmett G.m46AB+
Wesley R.m660-
# Import patient_data.xlsx, sheet 2: patients
patients <- read_excel("patient_data.xlsx", sheet = 2)

# Check the structure of patients
str(patients)

output:
tibble [4 × 4] (S3: tbl_df/tbl/data.frame)
 $ Name      : chr [1:4] "Hannah P." "Josh T." "Emmett G." "Wesley R."
 $ Sex       : chr [1:4] "f" "m" "m" "m"
 $ Age       : num [1:4] 25 33 46 66
 $ Blood type: chr [1:4] "B+" "B+" "AB+" "0-"

Importing Data from a JSON File​

JSON is short for JavaScript Object Notation. It is a free, universally accepted file format and method of exchanging data.

Before importing data from a JSON file, you need to load the rjson or the jsonlite package first. Both methods will be explored in the next sections.

rjson​

Install and load the rjson package:

install.packages("rjson")
library(rjson)

To convert a JSON object into R, you can use the fromJSON() function:

# Import the databases.json file: patient_data
patient_data <- fromJSON(file = "databases.json")

# Print out patient_data
patient_data

output:
$databases
$databases[[1]]
$databases[[1]]$name
[1] "Joey P."

$databases[[1]]$age
[1] "67"


$databases[[2]]
$databases[[2]]$name
[1] "Max C."

$databases[[2]]$age
[1] "49"


$databases[[3]]
$databases[[3]]$name
[1] "Selena M."

$databases[[3]]$age
[1] "36"

jsonlite​

Install and load the jsonlite package:

install.packages("jsonlite")
library(jsonlite)

The fromJSON() function is also available in jsonlite:

# Import the databases.json file: patient_data
patient_data <- fromJSON("databases.json")

# Print out patient_data
patient_data

output:
$databases
       name age
1   Joey P.  67
2    Max C.  49
3 Selena M.  36

As you can see, jsonlite utilized a different default data structure than rjson.

tip

Instead of a filename, you can also import URLs with fromJSON().

Importing Data from XML​

The Extensible Markup Language (short XML) is an extended markup language similar to HTML. To begin, install and load the XML as well as the xml2 package by typing the following commands into the R console:

install.packages("XML")
library(XML)

install.packages("xml2")
library(xml2)

read_xml()​

The XML data may be read via a path or URL to an XML file. For this, use the read_xml() function:

# Import the databases.xml file: patient_data
patient_data <- read_xml("databases.xml")

# Print out patient_data
patient_data

output:
{xml_document}
<databases>
[1] <database>\n  <name>Joey P.</name>\n  <age>67</age>\n</database>
[2] <database>\n  <name>Max C.</name>\n  <age>49</age>\n</database>
[3] <database>\n  <name>Selena M.</name>\n  <age>36</age>\n</database>

To make the data accessible, you can either parse the entire page or read elements separately.

xmlParse()​

Let's start by investigating the parsing option. Call xmlParse() and pass patient_data as an argument.

# Parse patient_data: patient_xml
patient_xml <- xmlParse(patient_data)

# Print out patient_xml
patient_xml

output:
<?xml version="1.0" encoding="UTF-8"?>
<databases>
  <database>
    <name>Joey P.</name>
    <age>67</age>
  </database>
  <database>
    <name>Max C.</name>
    <age>49</age>
  </database>
  <database>
    <name>Selena M.</name>
    <age>36</age>
  </database>
</databases>

Now, the contents resemble our source file.

Read Elements Separately​

If you want to read elements separately, you have the following options:

  • The xml_find_all() function may be used to get all items that share a tag. Any material included between the opening and closing tags is also provided.

    # Find all names in patient_data
    xml_find_all(patient_data, ".//name")

    output:
    {xml_nodeset (3)}
    [1] <name>Joey P.</name>
    [2] <name>Max C.</name>
    [3] <name>Selena M.</name>

  • If you merely require the content, you can use the xml_text(), xml_integer(), or xml_double() functions, depending on the underlying data type. In our case, the first option makes the most sense.

    # Find all names in patient_data (content only)
    xml_text(xml_find_all(patient_data, ".//name"))

    output:
    [1] "Joey P."   "Max C."    "Selena M."

Sources & Further Reading​