Ggplotly does not always make a legend correctly for where there are points and lines

Ggplotly does not always make a legend correctly for where there are points and lines. Sometimes it only makes the legend key based on either geom_point or geom_line, whichever is the first layer. Below is an example where it doesn't work. It seems to fix it if the data is arranged by the x variable. It'd be good if it worked for however the data was arranged.

library(dplyr)
library(ggplot2)
library(plotly)

#This doesn't work: the legend is based only on the first geom shape
plot_data <- iris %>% 
  as_tibble() %>%  
  janitor::clean_names() 

plot <- ggplot(data = plot_data, aes(petal_width, petal_length, col = species)) +
  geom_point() +
  geom_line()

plotly::ggplotly(plot)

plot_data <- iris %>% 
  as_tibble() %>%  
  janitor::clean_names() %>% 
  arrange(petal_width)

plot <- ggplot(data = plot_data, aes(petal_width, petal_length, col = species)) +
  geom_point() +
  geom_line()

plotly::ggplotly(plot)

Commencez par reproduire l’exemple R fourni avec plotly::ggplotly, en comparant les données iris non triées avec la version ordonnée par petal_width. L’issue est terminée lorsque la légende représente correctement les couches de points et de lignes, quel que soit l’ordre des lignes de données.