In this practical, you’ll learn how to import flat files using the readr package

Those kind of questions are optional

Before you start

To perform reproducible research it is a good practice to store the files in a standardized location. For example, you could take advantage of the RStudio projects and store data files in a sub-folder called data.

If you did not create an Rstudio project yet, create a RStudio project now.

Prepare your project’s folder

  1. Check that the project is active: the name you chose should appear on the top-right corner.

  2. Create a folder named data within your project’s folder. Use the Files pane in the lower right Rstudio panel or your favorite file browser.

  3. Download the file blood_fat.csv and place it in the data sub-folder you just created.

  4. Create a new Rmarkdown file, save it at the project root with a relevant name.

  5. Add a code chunk and with those lines to load the libraries. You don’t need to install the packages if those lines are working fine

library(dplyr)
library(readr)
  1. Don’t forget to run the chunk’s code to load the library during your interactive session

Warning

If you load the library only in the console and forget to place a chunk to load it, the knitting process will fail. Indeed, when you click on the knit button, the chunks are evaluated in a new and fresh environment.

Use readr to load your first file

Load the blood_fat file

Tip

the relative path can be safely built using "data/blood_fat.csv" if you followed the preliminary steps above, download the csv in a sub-folder data of a RStudio project

For example, you folder structure could be (depending on the picked names). Here:

  • RStudio project is Rworkshop
  • Rmarkdown document is practical02_import.Rmd
.
├── data
│   └── blood_fat.csv
├── practical03_import.Rmd
└── Rworkshop.Rproj
## Rows: 25 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): group
## dbl (4): id, weight, age, fat
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
## # A tibble: 25 × 5
##       id group weight   age   fat
##    <dbl> <chr>  <dbl> <dbl> <dbl>
##  1     1 A         84    46  354.
##  2     2 A         73    20  190.
##  3     3 A         65    52  406.
##  4     4 A         70    30  264.
##  5     5 A         76    57  452.
##  6     6 A         69    25  302.
##  7     7 A         63    28  288.
##  8     8 A         72    36  386.
##  9     9 A         79    57  402.
## 10    10 A         75    44  366.
## # … with 15 more rows

read_delim() execution is reporting the dimensions of the file, along with the guessed delimiter and data type of each columns

If we are happy with the guessed delimiter and the column names / types, we could silent this reporting.

Load again the same file, silencing the read_delim() message

The tibble

read_delim() loads the data as a tibble. The main advantage to use tibbles over a regular data frame is the printing.

  • Tibbles show some useful information such as the number of rows and columns:
    • Look at the top of the tibble and find the information “A tibble rows x cols”
    • How many rows are in the tibble?
  • The columns of a tibble report their type:
    • Look at the tibble header, the type of a columns is reported just below its name.
    • What is the type of the age column?

Actually, both age and id are integers, and should be read as such.

Read the blood_fat.csv specifying the data types of age and id as integers

Tip

In the col_types = c(....) you can use the columns bare names and either the long description to call the specific data type like col_integer() or the shortcut "i"
Read the blood_fat.csv specifying the data types of age and id as integers, skipping weight
One summarisation: compute the mean of both the age and weight per group
Copyright © Université du Luxembourg 2021. All rights reserved.