Problem with passing ggplot titles in a purrr loop (list-columns)

Question

I am using a list-columns approach with purr's map to generate series of plots with ggplot (I did read @aosmith's excellent posts both here printing ggplot with purrr map and on her blog https://aosmith.rbind.io/2018/08/20/automating-exploratory-plots/). However, where I am stuck is with how to pass individual titles to these plots right within the map loop. I know I can do it at a later stage with plotgrid and such, or with a mapply etc., but I was wondering why it doesn't work when I do it at the moment of plot generation with mutate.

Here is an example:

# Note that I am using dplyr 1.0.0 and R 4.0

library(tidyverse)

# Vector with names for the plots  

myvar <- c("cylinderA", "cylinderB", "cylinderC")
myvar <- set_names(myvar)

# Function to plot the plots
myPlot <- function(dataaa, namesss){
    ggplot(data = dataaa) +
        aes(x = disp, y = qsec) +
        geom_point() +
        labs(title = namesss)
}

# List-columns with map2 to iterate over split dataframe and the vector with names

mtcars %>% 
    group_by(cyl) %>% 
    nest() %>% 
    mutate(plots = map2(data, myvar, myPlot))

This produces an error pointing to a vector recycling problem, but my data list and myvar vector have the same length.

Error: Problem with `mutate()` input `plots`. x Input `plots` can't be recycled to size 1. ℹ Input `plots` is `map2(data, myvar, myPlot)`. ℹ Input `plots` must be size 1, not 3. ℹ The error occured in group 1: cyl = 4.

Also, when I use map2 on its own with the list-column and my vector with names, it works as expected...

test <- mtcars %>% 
    group_by(cyl) %>% 
    nest()

map2(test$data, myvar, myPlot) # This works

Any explanations of this behaviour and how to fix it within the purrr/dplyr approach would be much appreciated :-)

kJ

stefan stefan · Accepted Answer · 2020-06-11T12:21:33

Simply ungroup after the nest:

library(tidyverse)

# Vector with names for the plots  

myvar <- c("cylinderA", "cylinderB", "cylinderC")
myvar <- set_names(myvar)

# Function to plot the plots
myPlot <- function(dataaa, namesss){
  ggplot(data = dataaa) +
    aes(x = disp, y = qsec) +
    geom_point() +
    labs(title = namesss)
}

# List-columns with map2 to iterate over split dataframe and the vector with names

mtcars %>% 
  group_by(cyl) %>% 
  nest() %>% 
  ungroup() %>% 
  mutate(plots = map2(data, myvar, myPlot))
#> # A tibble: 3 x 3
#>     cyl data               plots 
#>   <dbl> <list>             <list>
#> 1     6 <tibble [7 x 10]>  <gg>  
#> 2     4 <tibble [11 x 10]> <gg>  
#> 3     8 <tibble [14 x 10]> <gg>

^{Created on 2020-06-11 by the reprex package (v0.3.0)}

Problem with passing ggplot titles in a purrr loop (list-columns)

1 Answers