Manual scales: not dropping and need exact order to merge with other aesthetics

[llrs]

Setting a manual color scale via scale_color_manual doesn't drop unused factors (this could be a feature and not a bug but it is not documented on the help page), and unless they are in the same exact order as other aesthetics the legends are not merged.

I wish there were some kind of check and reorder the scales to see if they match. I'm not sure how easy is to do this, but I think it would make ggplot2 slightly more user friendly.

I didn't find this mentioned previously on any other issue but I might have missed them.

library("ggplot2")
packageVersion("ggplot2")
#> [1] '3.3.6'
p <- ggplot(iris) +
  geom_point(aes(Petal.Length, Petal.Width, col = Species, shape = Species))

Manual scale

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue"))

Manual scale with extra values

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue", "Other" = "black"))

drop doesn’t seem to work with manual values

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue", "Other" = "black"),
                     drop = TRUE)

Manual scale with different order

p + scale_color_manual(values = c("versicolor" = "green", "virginica" = "blue", "setosa" = "red"))

The legends are not merged into a single one.

^{Created on 2022-06-03 by the reprex package (v2.0.1)}

[yutannihilation]

I forgot which change solved this, but the current dev version's behavior looks fine to me.

library("ggplot2")
packageVersion("ggplot2")
#> [1] '3.3.6.9000'

p <- ggplot(iris) +
  geom_point(aes(Petal.Length, Petal.Width, col = Species, shape = Species))

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue"))

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue", "Other" = "black"))

p + scale_color_manual(values = c("setosa" = "red", "versicolor" = "green", "virginica" = "blue", "Other" = "black"), drop = TRUE)

p + scale_color_manual(values = c("versicolor" = "green", "virginica" = "blue", "setosa" = "red"))

^{Created on 2022-06-04 by the reprex package (v2.0.1)}

[yutannihilation]

#4619