R plot：colors 不对应，图例不是预期的

Question

I have simplified my original code to explain my problem. My goal is to plot the values from the column "Value" of my data frame enter image description here . Each dot value would have the color mentionned in column "Color". The plot result is not what I want because:

• the colors don't correspond (is that a problem of "factor"?)
• I would like to have only one legend that displays: Blue color dot T1, Red color dot T2, green color dot T3

Would anyone know how to get that result? Thank you for your help !

DF <- DF_Total_Tidy[1:9,1:4]
colnames(DF)<-c("Time","Name","Value","Color")
DF[1:9,4]                                                                                  
<-c("blue","blue","blue","red","red","red","green","green","green")
DF$Name<- as.character(DF$Name)
DF[1:9,2]<-c("T1","T1","T1","T2","T2","T2","T3","T3","T3")
DF

Graph<- DF %>%
  ggplot(aes(x=Time, y=Value,fill=Name,color=Color)) +
  geom_point( size=4)
plot(Graph)

enter image description here

enter image description here

Answer 1

1. If you want to provide the color names via a column of your dataset you could use scale_color_identity(guide = "legend") where the guide argument is needed to add a legend (As a general rule I would recommend to not use something like DF$color .). One drawback of providing the colors is that you end up with the color names in the legend. To fix that you have to provide your desired labels via the labels argument, for which I use a named vector in my code below. Additionally your categories are ordered according to the color names. Hence, you probably have to use eg the breaks argument to set the right order of the categories in the legend.

2. To fix your second issue remove the fill aesthetic.

Using a minimal reprex based on mtcars I first reproduce your issue:

library(ggplot2)

mtcars$color <- with(mtcars, dplyr::case_when(cyl == 4 ~ "blue", cyl == 6 ~ "red", TRUE ~ "green"))
                     
ggplot(mtcars, aes(hp, mpg, color = color, fill = factor(cyl))) +
  geom_point()

Now, following the instructions outlined above we get:

ggplot(mtcars, aes(hp, mpg, color = color)) +
  geom_point() +
  scale_color_identity(
    guide = "legend",
    labels = setNames(unique(mtcars$cyl), unique(mtcars$color)),
    breaks = c("blue", "red", "green")
  )