Albert Rapp
Tidyselect helpers are neat little helper function that help you, well, select the right data. This can make your life easier when you want to either grab many different columns or even transform many columns using across() . For example, have you ever found yourself typing out repetitive calculations like this:
library(tidyverse)
penguins <- palmerpenguins::penguins
penguins |>
mutate(
bill_length_cm = bill_length_mm / 10,
bill_depth_cm = bill_depth_mm / 10,
flipper_length_cm = flipper_length_mm / 10
) |>
select(
-c(
bill_length_mm,
bill_depth_mm,
flipper_length_mm
)
)
#> # A tibble: 344 × 8
#> species island body_mass_g sex year bill_length_cm bill_depth_cm
#> <fct> <fct> <int> <fct> <int> <dbl> <dbl>
#> 1 Adelie Torgersen 3750 male 2007 3.91 1.87
#> 2 Adelie Torgersen 3800 female 2007 3.95 1.74
#> 3 Adelie Torgersen 3250 female 2007 4.03 1.8
#> 4 Adelie Torgersen NA <NA> 2007 NA NA
#> 5 Adelie Torgersen 3450 female 2007 3.67 1.93
#> 6 Adelie Torgersen 3650 male 2007 3.93 2.06
#> 7 Adelie Torgersen 3625 female 2007 3.89 1.78
#> 8 Adelie Torgersen 4675 male 2007 3.92 1.96
#> 9 Adelie Torgersen 3475 <NA> 2007 3.41 1.81
#> 10 Adelie Torgersen 4250 <NA> 2007 4.2 2.02
#> # ℹ 334 more rows
#> # ℹ 1 more variable: flipper_length_cm <dbl>
Here, we converted all the measurements that are given in millimeters (mm) to centimeters (cm). To do so, we had to
divide the mm numbers by 10 (yay, metric system 🥳) and
select all but the old columns (with units in mm)
Writing out all the new variable names is pretty repetitive and slightly annoying. At the least, it would be great if the last select() call could be simplified. And it turns out there is a way to do that, namely using tidyselect helpers. Let’s go through them all. And then, I’ll show you how you can use these tidyselect helpers to even avoid repetitive calculations in the first mutate() call.
Select consecutive columns
For now, though let’s just take our penguins data set as is and try to find ways to grab the right columns. Let’s take a look at the data to remind ourselves how it looks.
penguins
#> # A tibble: 344 × 8
#> species island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
#> <fct> <fct> <dbl> <dbl> <int> <int>
#> 1 Adelie Torgersen 39.1 18.7 181 3750
#> 2 Adelie Torgersen 39.5 17.4 186 3800
#> 3 Adelie Torgersen 40.3 18 195 3250
#> 4 Adelie Torgersen NA NA NA NA
#> 5 Adelie Torgersen 36.7 19.3 193 3450
#> 6 Adelie Torgersen 39.3 20.6 190 3650
#> 7 Adelie Torgersen 38.9 17.8 181 3625
#> 8 Adelie Torgersen 39.2 19.6 195 4675
#> 9 Adelie Torgersen 34.1 18.1 193 3475
#> 10 Adelie Torgersen 42 20.2 190 4250
#> # ℹ 334 more rows
#> # ℹ 2 more variables: sex <fct>, year <int>
Now, let’s grab consecutive columns with : . For example, we could grab the second to fourth column like this.
penguins |> select(2:4)
#> # A tibble: 344 × 3
#> island bill_length_mm bill_depth_mm
#> <fct> <dbl> <dbl>
#> 1 Torgersen 39.1 18.7
#> 2 Torgersen 39.5 17.4
#> 3 Torgersen 40.3 18
#> 4 Torgersen NA NA
#> 5 Torgersen 36.7 19.3
#> 6 Torgersen 39.3 20.6
#> 7 Torgersen 38.9 17.8
#> 8 Torgersen 39.2 19.6
#> 9 Torgersen 34.1 18.1
#> 10 Torgersen 42 20.2
#> # ℹ 334 more rows
That was pretty easy, right? This works exactly the same when you want to use column names instead of column numbers.
penguins |> select(island:bill_depth_mm)
#> # A tibble: 344 × 3
#> island bill_length_mm bill_depth_mm
#> <fct> <dbl> <dbl>
#> 1 Torgersen 39.1 18.7
#> 2 Torgersen 39.5 17.4
#> 3 Torgersen 40.3 18
#> 4 Torgersen NA NA
#> 5 Torgersen 36.7 19.3
#> 6 Torgersen 39.3 20.6
#> 7 Torgersen 38.9 17.8
#> 8 Torgersen 39.2 19.6
#> 9 Torgersen 34.1 18.1
#> 10 Torgersen 42 20.2
#> # ℹ 334 more rows
And the cool thing is you can still grab columns individually too. For example, let’s additionally grab the sex column. No problem. Just add this to select() .
penguins |> select(island:bill_depth_mm, sex)
#> # A tibble: 344 × 4
#> island bill_length_mm bill_depth_mm sex
#> <fct> <dbl> <dbl> <fct>
#> 1 Torgersen 39.1 18.7 male
#> 2 Torgersen 39.5 17.4 female
#> 3 Torgersen 40.3 18 female
#> 4 Torgersen NA NA <NA>
#> 5 Torgersen 36.7 19.3 female
#> 6 Torgersen 39.3 20.6 male
#> 7 Torgersen 38.9 17.8 female
#> 8 Torgersen 39.2 19.6 male
#> 9 Torgersen 34.1 18.1 <NA>
#> 10 Torgersen 42 20.2 <NA>
#> # ℹ 334 more rows
That’s the beauty of all these selection helpers. You can combine them as much as you want.
Grab columns based on data type
The : operator assumes that we want columns that come in a consecutive order. But sometimes we have so many columns that we are only interested in particular columns based on their data types. Maybe we want to calculate means with these columns. That would be hard to do with columns that don’t have numbers in them.
Thankfully, we have a tidyselect helper that grabs columns based on data types. But it requires another partner function that evaluates the data type. Check it out.
penguins |>
select(
where(is.numeric)
)
#> # A tibble: 344 × 5
#> bill_length_mm bill_depth_mm flipper_length_mm body_mass_g year
#> <dbl> <dbl> <int> <int> <int>
#> 1 39.1 18.7 181 3750 2007
#> 2 39.5 17.4 186 3800 2007
#> 3 40.3 18 195 3250 2007
#> 4 NA NA NA NA 2007
#> 5 36.7 19.3 193 3450 2007
#> 6 39.3 20.6 190 3650 2007
#> 7 38.9 17.8 181 3625 2007
#> 8 39.2 19.6 195 4675 2007
#> 9 34.1 18.1 193 3475 2007
#> 10 42 20.2 190 4250 2007
#> # ℹ 334 more rows
Here, we have partnered up the tidyselect helper where() with the is.numeric function. The latter function checks whether a column is numeric and returns TRUE or FALSE for each column accordingly. Then, where() grabs those responses from is.numeric and grabs only those columns where is.numeric returned TRUE.
Notice that you don’t actually call is.numeric in your code. You only do where(is.numeric) and not where(is.numeric()) . That’s because you partner up the function is.numeric with the where function. This way, the where function just knows the name of its partner (in this case is.numeric ) and can let its partner do the checking whenever where feels like it.
Grab columns based on name
Of course, sometimes you don’t even have to think this complicated. Sometimes, you actually know which numeric columns you want to grab. For example, in our first example we wanted to transform the millimeter columns. So, we could just grab the columns that have a mm in their column name. That works via contains() .
penguins |>
select(
contains("mm")
)
#> # A tibble: 344 × 3
#> bill_length_mm bill_depth_mm flipper_length_mm
#> <dbl> <dbl> <int>
#> 1 39.1 18.7 181
#> 2 39.5 17.4 186
#> 3 40.3 18 195
#> 4 NA NA NA
#> 5 36.7 19.3 193
#> 6 39.3 20.6 190
#> 7 38.9 17.8 181
#> 8 39.2 19.6 195
#> 9 34.1 18.1 193
#> 10 42 20.2 190
#> # ℹ 334 more rows
In this particular case, we could even use the ends_with() helper. Using this alternative can sometimes be helpful to avoid accidentally picking a column that for some other reason has mm in it.
penguins |>
select(
ends_with("mm")
)
#> # A tibble: 344 × 3
#> bill_length_mm bill_depth_mm flipper_length_mm
#> <dbl> <dbl> <int>
#> 1 39.1 18.7 181
#> 2 39.5 17.4 186
#> 3 40.3 18 195
#> 4 NA NA NA
#> 5 36.7 19.3 193
#> 6 39.3 20.6 190
#> 7 38.9 17.8 181
#> 8 39.2 19.6 195
#> 9 34.1 18.1 193
#> 10 42 20.2 190
#> # ℹ 334 more rows
Similarly, there’s start_with() . For example, we can grab the bill-related columns that way.
penguins |>
select(
starts_with("bill")
)
#> # A tibble: 344 × 2
#> bill_length_mm bill_depth_mm
#> <dbl> <dbl>
#> 1 39.1 18.7
#> 2 39.5 17.4
#> 3 40.3 18
#> 4 NA NA
#> 5 36.7 19.3
#> 6 39.3 20.6
#> 7 38.9 17.8
#> 8 39.2 19.6
#> 9 34.1 18.1
#> 10 42 20.2
#> # ℹ 334 more rows
Simplify our first example
Now we have all the tools we need to simplify our first example. Recall that the code looked like this:
penguins |>
mutate(
bill_length_cm = bill_length_mm / 10,
bill_depth_cm = bill_depth_mm / 10,
flipper_length_cm = flipper_length_mm / 10,
) |>
select(
-c(
bill_length_mm, bill_depth_mm,
flipper_length_mm
)
)
#> # A tibble: 344 × 8
#> species island body_mass_g sex year bill_length_cm bill_depth_cm
#> <fct> <fct> <int> <fct> <int> <dbl> <dbl>
#> 1 Adelie Torgersen 3750 male 2007 3.91 1.87
#> 2 Adelie Torgersen 3800 female 2007 3.95 1.74
#> 3 Adelie Torgersen 3250 female 2007 4.03 1.8
#> 4 Adelie Torgersen NA <NA> 2007 NA NA
#> 5 Adelie Torgersen 3450 female 2007 3.67 1.93
#> 6 Adelie Torgersen 3650 male 2007 3.93 2.06
#> 7 Adelie Torgersen 3625 female 2007 3.89 1.78
#> 8 Adelie Torgersen 4675 male 2007 3.92 1.96
#> 9 Adelie Torgersen 3475 <NA> 2007 3.41 1.81
#> 10 Adelie Torgersen 4250 <NA> 2007 4.2 2.02
#> # ℹ 334 more rows
#> # ℹ 1 more variable: flipper_length_cm <dbl>
We no longer have to type out all the mm column names to remove them (with - ). We can just use ends_with() .
penguins |>
mutate(
bill_length_cm = bill_length_mm / 10,
bill_depth_cm = bill_depth_mm / 10,
flipper_length_cm = flipper_length_mm / 10
) |>
select(-ends_with("mm"))
#> # A tibble: 344 × 8
#> species island body_mass_g sex year bill_length_cm bill_depth_cm
#> <fct> <fct> <int> <fct> <int> <dbl> <dbl>
#> 1 Adelie Torgersen 3750 male 2007 3.91 1.87
#> 2 Adelie Torgersen 3800 female 2007 3.95 1.74
#> 3 Adelie Torgersen 3250 female 2007 4.03 1.8
#> 4 Adelie Torgersen NA <NA> 2007 NA NA
#> 5 Adelie Torgersen 3450 female 2007 3.67 1.93
#> 6 Adelie Torgersen 3650 male 2007 3.93 2.06
#> 7 Adelie Torgersen 3625 female 2007 3.89 1.78
#> 8 Adelie Torgersen 4675 male 2007 3.92 1.96
#> 9 Adelie Torgersen 3475 <NA> 2007 3.41 1.81
#> 10 Adelie Torgersen 4250 <NA> 2007 4.2 2.02
#> # ℹ 334 more rows
#> # ℹ 1 more variable: flipper_length_cm <dbl>
Ahh that’s better better. Much less typing.
Tidyselect helpers everywhere
Now that we have that covered, let me show you how to simplify the repetitive mutate() call too. We can actually use tidyselect helpers there as well. After all, we could tell mutate() to take all the columns that end in mm and divide their values by 10. To do so, we first define a function that does the tricky math part.
divide_by_ten <- function(x) x / 10
And then we combine mutate() with across() to apply divide_by_ten across desired columns.
penguins |>
mutate(
across(
.cols = ends_with("mm"), # <- tidyselect helper here
.fns = divide_by_ten
)
)
#> # A tibble: 344 × 8
#> species island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
#> <fct> <fct> <dbl> <dbl> <dbl> <int>
#> 1 Adelie Torgersen 3.91 1.87 18.1 3750
#> 2 Adelie Torgersen 3.95 1.74 18.6 3800
#> 3 Adelie Torgersen 4.03 1.8 19.5 3250
#> 4 Adelie Torgersen NA NA NA NA
#> 5 Adelie Torgersen 3.67 1.93 19.3 3450
#> 6 Adelie Torgersen 3.93 2.06 19 3650
#> 7 Adelie Torgersen 3.89 1.78 18.1 3625
#> 8 Adelie Torgersen 3.92 1.96 19.5 4675
#> 9 Adelie Torgersen 3.41 1.81 19.3 3475
#> 10 Adelie Torgersen 4.2 2.02 19 4250
#> # ℹ 334 more rows
#> # ℹ 2 more variables: sex <fct>, year <int>
Here, this overwrites the existing columns. But these still end in mm . To fix that, we can use .names argument in across() to fix that. We use str_remove() on the previous column names (encoded as .col ) and append the _cm ending ourselves.
penguins |>
mutate(
across(
.cols = ends_with("mm"),
.fns = divide_by_ten,
.names = '{str_remove(.col, "_mm")}_cm'
)
) |>
select(-ends_with("mm")) # remove old columns here
#> # A tibble: 344 × 8
#> species island body_mass_g sex year bill_length_cm bill_depth_cm
#> <fct> <fct> <int> <fct> <int> <dbl> <dbl>
#> 1 Adelie Torgersen 3750 male 2007 3.91 1.87
#> 2 Adelie Torgersen 3800 female 2007 3.95 1.74
#> 3 Adelie Torgersen 3250 female 2007 4.03 1.8
#> 4 Adelie Torgersen NA <NA> 2007 NA NA
#> 5 Adelie Torgersen 3450 female 2007 3.67 1.93
#> 6 Adelie Torgersen 3650 male 2007 3.93 2.06
#> 7 Adelie Torgersen 3625 female 2007 3.89 1.78
#> 8 Adelie Torgersen 4675 male 2007 3.92 1.96
#> 9 Adelie Torgersen 3475 <NA> 2007 3.41 1.81
#> 10 Adelie Torgersen 4250 <NA> 2007 4.2 2.02
#> # ℹ 334 more rows
#> # ℹ 1 more variable: flipper_length_cm <dbl>
If you found those last steps a bit confusing, don’t worry about it too much. The point here simply is that tidyselect helpers can be used in all kinds of situations. And with that we have learned another tool for our data analysis journey. 🎉 Hope you’ll get lots of value out of this one. Enjoy the rest of your day and see you next time 👋
You need to be signed-in to comment on this post. Login.