ggplot reorders categorical variables

Question

I am having trouble figuring out why ggplot reorders my categorical variables

xaxis = c('80','90','100')
test = data.frame(x = xaxis, y = c(1,2,3))
ggplot(test, aes(x=x,y=y)) + geom_point()

I found online that it has something to do with the factor levels and the following code fixes my problem.

xaxis = c('80','90','100')
xaxis = factor(xaxis,levels=xaxis)
test = data.frame(x = xaxis, y = c(1,2,3))
ggplot(test, aes(x=x,y=y)) + geom_point()

But if you go back to the original code.

class(xaxis)
[1] "character"

Its simply a character vector and I don't see any innate ordering. Can someone explain what is happening here please? Do I always have to change my x variable into a factor for ggplot to respect my sequence?

Answer 1

sort(xaxis)
[1] "100" "80"  "90"

Sorting of character vectors is done by a character by character basis - ie it doesn't understand the numerical context of the data.

ggplot2 will convert character variables to factors and by default factors sort their levels:

factor(xaxis)
[1] 80  90  100
Levels: 100 80 90