Why cant i use vector in scale fill manual?-CodePudding

I Want to create a graph that uses scale fill manual to fill colours of graph.

G3 <- ggplot(preTMM, aes(Var2, value, fill=group)) 
  geom_boxplot() 
  Gtheme2 
  xlab("") 
  ylab("Log-CPM") 
  scale_fill_manual(values=c("pCDNA"="grey59","M"= "steelblue3"),
                    breaks=c("pCDNA","M")) 
  guides(fill=FALSE)

This works fine, but if i want to add it via a vector that was made previously;

DGEGroup1 <- "pCDNA"
DGEGroup2 <- "M" 
DGEGroup1Col <- "grey59"
DGEGroup2Col <- "steelblue3"

it doesnt work.. what am I doing wrong.

G3 <- ggplot(preTMM, aes(Var2, value, fill=group)) 
  geom_boxplot() 
  Gtheme2 
  xlab("") 
  ylab("Log-CPM") 
scale_fill_manual(values=c(DGEGroup1=DGEGroup1Col,DGEGroup2=DGEGroup1Col),
                    breaks=c(DGEGroup1,DGEGroup2)) 
  guides(fill=FALSE)

I want to do this so all the graphs can be controlled by one colour vector if that makes sense?

Thanks in advance for help!

CodePudding user response：

That syntax for a named vector doesn't work because R functions (including c()) don't require quotes for argument names:

x = "a"
y = "b"
c(x = y)
#   x 
# "b" 
## See how the name is literally `"x"`, not `"a"`

This is a Good Thing. It means that functions that have an argument named x don't break when you also have an object named x.

For your purpose, a good fix is to use setNames to name the vector. (You could also use names(vector) <- ... but setNames lets you keep it as a one-liner with an anonymous vector, much like you already have it).

values = setNames(
  c(DGEGroup1Col, DGEGroup2Col),
  nm = c(DGEGroup1, DGEGroup2)
)

However, if you're doing this for multiple plots I'd strongly suggest pulling out the definition of that vector earlier in you code so you're not repeating this syntax.

CodePudding user response：

You haven't supplied a reproducible example, but as Gregor Thomas points out, this just isn't how you create named vectors in R. Here's an example using the built-in iris dataset that shows how you could use your existing variables to get the intended results.

Firstly, I'll assign the group names and colour names to match your set-up:

Group1 <- "setosa"
Group2 <- "versicolor"
Group3 <- "virginica"

Group1col  <- "red"
Group2col  <- "green"
Group3col  <- "blue"

To easily create a named vector, we can use setNames:

setNames(c(Group1col, Group2col, Group3col), c(Group1, Group2, Group3))
#>     setosa versicolor  virginica 
#>      "red"    "green"     "blue"

So to use the named variables in your plot, you could do:

ggplot(iris, aes(Sepal.Width, Petal.Length, colour = Species))  
  geom_point()  
  scale_color_manual(values = setNames(c(Group1col, Group2col, Group3col), 
                                       c(Group1, Group2, Group3)))