R ggplot2: legend should be discrete and not continuous

问题

I have the following data:

benchmark   mispredpenal    IPC     pred
ammp        1               1.0589  2lev
ammp        5               1.0450  2lev
...

and use the following command:

ggplot(IPC, aes(x = benchmark, y = IPC, group=mispredpenal, colour=mispredpenal)) + 
  geom_point() + geom_line()

Everything looks like it should, but I would like the legend to be discrete, and not the continuous (gradient). How should I do this?

Edit: Misprediction is either 1, 5, 9, 13 or 17.

回答1:

You want the variable mispredpenal to be a factor in that case:

ggplot(IPC, aes(x = benchmark, y = IPC, group=factor(mispredpenal), colour=factor(mispredpenal))) + 
  geom_point() + geom_line()

来源：https://stackoverflow.com/questions/20301922/r-ggplot2-legend-should-be-discrete-and-not-continuous