Specific colours on stacked Bars in R with ggplot2

Question

this is my code in R:

prod <- read.csv("/tmp/pepper.csv",header=T,sep="\t")
png("/tmp/image.png", width=1000, height=1000)

datm <- melt(cbind(prod,ind = rownames(prod)),is.vars = c('ind'))

ggplot(datm,aes(x = codigo_inver,y = value,fill = factor(variable))) + 
  geom_bar(stat='identity', position = "fill" ) +
  coord_flip()+
  scale_colour_manual(values = c("red","orange", "green"))
dev.off()

Data set is:

green   orange  red codigo_inver
48.40   30.22   21.38   7_7726-14
32.31   28.18   39.51   8_7726-14
46.74   30.13   23.13   9_7577-4
55.13   32.80   12.06   21_7562-4
51.30   30.76   17.94   28_7614-1
40.65   37.75   21.60   30_7094-2

and it produce:

But the colors mismatch with

values = c("red","orange", "green")

In the picture, red should be green , green should be orange and blue should be red.

Colours must match with the column name data, red is red, orange is orange and green is green.

How can I match this colours with original column name color?

thanks in advance!

Answer 1

Get the colours from the levels of the factor, and use scale_fill_manual:

ggplot(datm,aes(x = codigo_inver,y = value,fill = factor(variable))) + 
  geom_bar(stat='identity', position = "fill" ) +
  coord_flip()+
  scale_fill_manual(values = levels(factor(datm$variable)))

however the legend now looks a bit obvious...

Answer 2

As you use fill to assign the colors, you should also use the corresponding scale:

scale_fill_manual(values = c("red","orange", "green"))

now it should work