CODEWITHSUNDEEP
rfunctiondplyrtidyevalquasiquotes
daaronr
daaronr

Reputation: 517

tidyr use glue strings later in function

Tidy eval now supports glue strings

So this works great:

my_summarise5 <- function(data, mean_var ) {
  data %>% 
    mutate(
      "mean_{{mean_var}}" := mean({{ mean_var }}), 
    ) 
}

mtcars %>% my_summarise5(cyl)

But then

my_summarise5 <- function(data, mean_var ) {
  data %>% 
    mutate(
      "mean_{{mean_var}}" := mean({{ mean_var }}), 
    "mean_{{mean_var}}_plusone" :=  "mean_{{mean_var}}"+1
    )
}

mtcars %>% my_summarise5(cyl)

Throws

Error: Problem with `mutate()` input `mean_cyl_plusone`.
x non-numeric argument to binary

Would some 'paste' or 'glue' thing in the "mean_{{mean_var}}_plusone" := "mean_{{mean_var}}"+1 part fix this?

Note this is obviously not a useful case, its a MWE for the syntax. I actually want to define two new columns with different names, one which uses the other ... otherwise I have to repeat and it also gets messy.

Upvotes: 4

Views: 308

Answers (2)

G. Grothendieck
G. Grothendieck

Reputation: 269301

Use across:

my_summarise5 <- function(data, mean_var ) {
  data %>% 
    mutate(
      "mean_{{mean_var}}" := mean({{ mean_var }}),
      across(last_col(), ~.+1, .names = "{col}_plusone")
    )
}

mtcars %>% my_summarise5(cyl) %>% head

giving:

                   mpg cyl disp  hp drat    wt  qsec vs am gear carb mean_cyl mean_cyl_plusone
Mazda RX4         21.0   6  160 110 3.90 2.620 16.46  0  1    4    4   6.1875           7.1875
Mazda RX4 Wag     21.0   6  160 110 3.90 2.875 17.02  0  1    4    4   6.1875           7.1875
Datsun 710        22.8   4  108  93 3.85 2.320 18.61  1  1    4    1   6.1875           7.1875
Hornet 4 Drive    21.4   6  258 110 3.08 3.215 19.44  1  0    3    1   6.1875           7.1875
Hornet Sportabout 18.7   8  360 175 3.15 3.440 17.02  0  0    3    2   6.1875           7.1875
Valiant           18.1   6  225 105 2.76 3.460 20.22  1  0    3    1   6.1875           7.1875

Upvotes: 4

dash2
dash2

Reputation: 2262

I love the tidyverse, but using base R can have a big payoff, by making your code readable and maintainable.

my_summarise5 <- function (data, mean_var) {
  mean_name <- paste0("mean_", mean_var)
  plus_one_name <- paste0(mean_name, "_plus_one")

  data[[mean_name]] <- mean(data[[mean_var]])
  data[[plus_one_name]] <- data[[mean_name]] + 1

  data
}

Anybody who knows R can understand the above.

tmp <- data.frame(a = 1:5)
my_summarise5(tmp, "a")
##   a mean_a mean_a_plus_one
## 1 1      3               4
## 2 2      3               4
## 3 3      3               4
## 4 4      3               4
## 5 5      3               4

If you want to pass a bare symbol as mean_var, then just add this line to the start of the function:

mean_var <- as.character(substitute(mean_var))

This isn't perfect, because it'll break with deeply nested functions. So you could mix a little tidyverse in by doing (something like) mean_var <- enquo(mean_var). But I still would prefer the simple data manipulation to stuff with mutate and across.

Upvotes: 1

Related Questions