new_scale_colour does not work with a scale_* function that includes a guide argument

[davidhodge931]

library(palmerpenguins)
library(tidyverse)
#> Warning: package 'purrr' was built under R version 4.3.1
library(ggnewscale)
#> Warning: package 'ggnewscale' was built under R version 4.3.1

d <- penguins |>
  group_by(sex, species) |>
  summarise(across(flipper_length_mm, \(x) mean(x, na.rm = TRUE))) |>
  tidyr::drop_na(sex) |>
  mutate(species = stringr::str_to_sentence(species))
#> `summarise()` has grouped output by 'sex'. You can override using the `.groups`
#> argument.


p <- d |>
  ggplot(aes(x = flipper_length_mm, colour = species, fill = species, y = sex)) +
  geom_col(position = "dodge", width = 0.75, alpha = 0.8) +
  scale_colour_manual(values = c("lightgrey", "darkblue", "orange"),
                      guide = ggplot2::guide_legend(reverse = T, title.position = "top"),
                      aesthetics = c("colour", "fill"))

##########################
#does not work with a guide
##########################
p +
  ggnewscale::new_scale_colour()
#> Error in scale$guide == "none": comparison (==) is possible only for atomic and list types

p <- d |>
  ggplot(aes(x = flipper_length_mm, colour = species, fill = species, y = sex)) +
  geom_col(position = "dodge", width = 0.75, alpha = 0.8) +
  scale_colour_manual(values = c("lightgrey", "darkblue", "orange"),
                      # guide = ggplot2::guide_legend(reverse = T, title.position = "top"),
                      aesthetics = c("colour", "fill"))

##########################
#works without a guide
##########################
p +
  ggnewscale::new_scale_colour()

^{Created on 2023-09-05 with reprex v2.0.2}

[davidhodge931]

This is using the DEV version of ggplot2 and ggnewscale

Fantastic package btw :)

[davidhodge931]

If I put the guides arguments as a seperate layer instead of from within the scale function, then it allows the new_scale_color layer but creates problems with the legend by ignoring the initial colour scale guide arguments.

library(tidyverse)
#> Warning: package 'purrr' was built under R version 4.3.1
#> Warning: package 'dplyr' was built under R version 4.3.1
library(palmerpenguins)
library(ggnewscale)

d <- penguins |>
  group_by(sex, species) |>
  summarise(across(flipper_length_mm, \(x) mean(x, na.rm = TRUE))) |>
  tidyr::drop_na(sex) |>
  mutate(species = stringr::str_to_sentence(species))
#> `summarise()` has grouped output by 'sex'. You can override using the `.groups`
#> argument.

d |>
  ggplot(aes(y = flipper_length_mm, x = sex, colour = species, fill = species)) +
  geom_col(position = "dodge", width = 0.75, alpha = 0.8) +
  scale_colour_manual(values = c("lightgrey", "darkblue", "orange"),
                      aesthetics = c("colour", "fill")) +
  guides(colour = ggplot2::guide_legend(reverse = TRUE, title.position = "top"),
         fill = ggplot2::guide_legend(reverse = TRUE, title.position = "top")) +
  new_scale_color()

^{Created on 2023-09-06 with reprex v2.0.2}

[davidhodge931]

I assume that this relates to this??
#57

[eliocamp]

I assume that this relates to this?? #57

Yes, that PR fixed this issue.