'unused argument' error when using `purrr::pwalk` to save ggplots from list column in nested dataframe

I’m trying to save out ggplots from a list column in a nested dataframe with the nesting variable in the filename so that I know which plot corresponds to which data.

I’m using an approach which worked for me saving csv files from nested dataframes, but doesn’t work with ggplot objects.

data("mtcars")
temp_dir <- tempfile()
dir.create(temp_dir)
library(tidyr) # for nest function

plot_fun <- function(x){
  p <- ggplot(data = x, aes(x = wt, y = mpg)) + 
    geom_point()
  return(p)
}

cyl_mtcars <- mtcars %>% 
  nest(data = -cyl) %>% 
  mutate(plots = map(.x = data, .f = plot_fun)
  )
head(cyl_mtcars)

Which produces

# A tibble: 3 × 3
    cyl data               plots 
  <dbl> <list>             <list>
1     6 <tibble [7 × 10]>  <gg>  
2     4 <tibble [11 × 10]> <gg>  
3     8 <tibble [14 × 10]> <gg>

But when I try to save the figures out

cyl_mtcars %>% 
  pwalk(
    function(cyl, plots) ggsave(filename = file.path(temp_dir, glue("{cyl}.png")), plot = plots)
    )

I get this error

Error in `pmap()` at purrr/R/pmap.R:148:2:
ℹ In index: 1.
Caused by error in `.f()`:
! unused argument (data = .l[[2]][[i]])
Run `rlang::last_trace()` to see where the error occurred.```

>Solution :

The issue is that using pwalk or pmap all columns are passed to the function. And as your function has only two arguments you get an unused argument error as you nested data has three columns. To fix that, you could add ... to the function arguments:

temp_dir <- tempfile()
dir.create(temp_dir)

library(tidyverse)

plot_fun <- function(x) {
  ggplot(data = x, aes(x = wt, y = mpg)) +
    geom_point()
}

cyl_mtcars <- mtcars %>%
  nest(data = -cyl) %>%
  mutate(plots = map(.x = data, .f = plot_fun))

cyl_mtcars %>% 
  pwalk(
    function(cyl, plots, ...) {
      ggsave(filename = file.path(temp_dir, glue::glue("{cyl}.png")), plot = plots)
    }
  )
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image
#> Saving 7 x 5 in image