library(tidyverse)
theme_set(theme_classic())First, let’s talk about variables. There are two kinds of variables we need to know about.
env-variables are variables that exist in our environment. They’re
what you typically think of when you think of the word variable. In R,
you create them with the assignment operator (<-):
name <- "Kelly"
numbers <- 1:3
ls()## [1] "name" "numbers"
Besides variables that you create, some packages include datasets that
are also available as env-variables. iris is a dataframe that’s
included in base R.
head(iris)## Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1 5.1 3.5 1.4 0.2 setosa
## 2 4.9 3.0 1.4 0.2 setosa
## 3 4.7 3.2 1.3 0.2 setosa
## 4 4.6 3.1 1.5 0.2 setosa
## 5 5.0 3.6 1.4 0.2 setosa
## 6 5.4 3.9 1.7 0.4 setosa
data-variables are columns of dataframes. They aren’t in our
environment – they live in their dataframes.
For example, Species is a data-variable in iris. The following
code doesn’t work:
print(Species)## Error in print(Species): object 'Species' not found
In base R, we have to tell R which env-variables to access
data-variables from using $:
iris[iris$Species == "setosa" & iris$Sepal.Length > 5.5, ]## Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 15 5.8 4.0 1.2 0.2 setosa
## 16 5.7 4.4 1.5 0.4 setosa
## 19 5.7 3.8 1.7 0.3 setosa
The tidyverse way allows us to avoid repeating the name of the dataframe:
iris %>% filter(Species == "setosa", Sepal.Length > 5.5)## Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1 5.8 4.0 1.2 0.2 setosa
## 2 5.7 4.4 1.5 0.4 setosa
## 3 5.7 3.8 1.7 0.3 setosa
There’s some magic going on under the hood of dplyr’s filter function
that lets us use Species and Sepal.Length, which are
data-variables, as if they were env-variables. Normally, R would
complain that it can’t find Species.
Tidy evaluation is used in many tidyverse functions to delay evaluating the variables until later on inside the function code. This makes it slightly harder to use tidyverse functions inside our own functions, but it’s worth learning to write more reusable code!
Let’s say you’re summarizing your data:
iris %>%
group_by(Species) %>%
summarize(
n = n(),
min = min(Sepal.Length),
max = max(Sepal.Length)
)## # A tibble: 3 x 4
## Species n min max
## <fct> <int> <dbl> <dbl>
## 1 setosa 50 4.3 5.8
## 2 versicolor 50 4.9 7
## 3 virginica 50 4.9 7.9
And you realize you’d like to also summarize the Sepal.Width column.
You could copy-paste your code and change both instances of
Sepal.Length to Sepal.Width:
iris %>%
group_by(Species) %>%
summarize(
n = n(),
min = min(Sepal.Width),
max = max(Sepal.Width)
)## # A tibble: 3 x 4
## Species n min max
## <fct> <int> <dbl> <dbl>
## 1 setosa 50 2.3 4.4
## 2 versicolor 50 2 3.4
## 3 virginica 50 2.2 3.8
But now you have repetitive code. If you later decide you want to also calculate the mean of those columns, you’ll have to modify your code in two places – or more if you kept on copy-pasting for more columns.
Instead, let’s try writing a function that will summarize any column name we want.
var_summary <- function(data, var) {
data %>%
group_by(Species) %>%
summarize(n = n(),
min = min(var),
max = max(var))
}
iris %>%
var_summary(Sepal.Length)## Error: object 'Sepal.Length' not found
That didn’t quite work because Sepal.Length is a data-variable that
doesn’t exist in our environment, so R can’t find it.
We need to tell R to wait to evaluate the var argument until it gets
to the summarize function.
We can do that by capturing the variable with the enquo function to
create a quosure (quoted closure). A quosure captures a variable
or expression so we can evaluate it later. Think of it like freezing in
place. Using a special operator !! (pronounced bang-bang), we’ll
tell R to wait to evaluate our quosure until it’s inside the summarize
function.
var_summary <- function(data, var) {
var <- enquo(var)
data %>%
group_by(Species) %>%
summarize(n = n(),
min = min(!!var),
max = max(!!var))
}
iris %>%
var_summary(Sepal.Length)## # A tibble: 3 x 4
## Species n min max
## <fct> <int> <dbl> <dbl>
## 1 setosa 50 4.3 5.8
## 2 versicolor 50 4.9 7
## 3 virginica 50 4.9 7.9
If you don’t need to do anything with the quosure before unquoting it,
you can use !!enquo all at once like this:
var_summary <- function(data, var) {
data %>%
group_by(Species) %>%
summarize(n = n(),
min = min(!!enquo(var)),
max = max(!!enquo(var)))
}
iris %>%
var_summary(Sepal.Length)## # A tibble: 3 x 4
## Species n min max
## <fct> <int> <dbl> <dbl>
## 1 setosa 50 4.3 5.8
## 2 versicolor 50 4.9 7
## 3 virginica 50 4.9 7.9
This pattern of using !!enquo is really common, so the tidyverse
developers created a new operator called embracing {{ }} to do them
both at the same time:
var_summary <- function(data, var) {
data %>%
group_by(Species) %>%
summarize(n = n(),
min = min({{ var }}),
max = max({{ var }})
)
}Most of the time, embracing is what you’ll want to use.
You can also pass more complicated expressions as arguments to functions.
filter2 <- function(data, comparison1, comparison2) {
data %>%
filter({{ comparison1 }}, {{ comparison2 }})
}
iris %>% filter2(Sepal.Length > Sepal.Width * 2.5, Species == "versicolor")## Sepal.Length Sepal.Width Petal.Length Petal.Width Species
## 1 6.0 2.2 4.0 1.0 versicolor
## 2 6.2 2.2 4.5 1.5 versicolor
## 3 6.3 2.5 4.9 1.5 versicolor
## 4 6.3 2.3 4.4 1.3 versicolor
You can pass any number of arguments to other functions with ...
summarize_vars <- function(.data, ...) {
.data %>%
group_by(Species) %>%
summarize_at(vars(...), mean)
}
iris %>% summarize_vars(Sepal.Length, Sepal.Width, Petal.Width)## # A tibble: 3 x 4
## Species Sepal.Length Sepal.Width Petal.Width
## <fct> <dbl> <dbl> <dbl>
## 1 setosa 5.01 3.43 0.246
## 2 versicolor 5.94 2.77 1.33
## 3 virginica 6.59 2.97 2.03
summarize_vars <- function(.data, ...) {
.data %>%
group_by(...) %>%
summarize_at(vars(-group_cols()), mean)
}
iris %>% summarize_vars(Species)## # A tibble: 3 x 5
## Species Sepal.Length Sepal.Width Petal.Length Petal.Width
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 setosa 5.01 3.43 1.46 0.246
## 2 versicolor 5.94 2.77 4.26 1.33
## 3 virginica 6.59 2.97 5.55 2.03
Note:
- You don’t need to use
!!enquoor{{ }}when using...– it just works! - The triple-dots
...has to be the last argument in the function signature. - It’s considered good practice to name your other arguments beginning
with a
.so users don’t accidentally match them to other variables forwarded by...(more details here).
# replace the X's to make this code work
calc_ratio <- function(data, var1, var2) {
XXXXXX %>%
mutate(ratio = XXXXXXX / XXXXXXX)
}
# example:
#iris %>% calc_ratio(Sepal.Length, Sepal.Width)
#iris %>% calc_ratio(Petal.Width, Petal.Length)Here’s an example histogram where x is Sepal.Length:
iris %>%
ggplot(aes(x = Sepal.Length, fill = Species)) +
geom_histogram()## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
Create a function that takes a column to use for x:
plot_hist <- function(data, x_col) {
# write code here
}
plot_hist(iris, Sepal.Width)## NULL
plot_hist(iris, Petal.Length)## NULL
# write function here
# example:
#iris %>% plot_box(x = Species, y = Petal.Width, fill = Species)
#iris %>% plot_box(x = Species, y = Sepal.Length, color = Species)How strong is your Google-foo?
calc_mean <- function(data, var) {
data %>%
summarize(XXXXXXXXX= mean({{ var }}))
}
# example:
#iris %>% calc_mean(Sepal.Length)
#> mean_Sepal.Length
#> 1 5.843333- Vignette on tidy eval with dplyr (this document adapts parts of it): https://dplyr.tidyverse.org/dev/articles/programming.html
- Section on tidy eval in The Tidynomicon: http://tidynomicon.tech/nse.html#what-is-tidy-evaluation
- Tidy Eval book (in progress): https://tidyeval.tidyverse.org/
- Metaprogramming chapters of Advanced R: https://adv-r.hadley.nz/metaprogramming.html