CODEWITHSUNDEEP
rggplot2
sam
sam

Reputation: 541

Is there a "geom_donothing" function in ggplot?

When I use ggplot, I generally format my code as such:

ggplot() +
   geom_col() +
   geom_line() +
   scale_y_continuous()

This format neatly organizes the components of the plot, and RStudio lets you run all the lines by hitting Ctrl+Enter. If I ever want to remove a component of the plot, I can simply comment that line out, and RStudio will ignore that line.

But if I comment out the last line,

ggplot() +
   geom_col() +
   geom_line() +
   #scale_y_continuous()

I will get an error, because I did not remove the + after geom_line(). It is obviously easy to remove the + sign, but then I have add it back when I un-comment the last line.

It's not a huge problem, but in the process of making a plot, I inevitably forget to add or remove the + at least once. It is a bigger problem when working on a laptop with a small screen, where I cannot necessarily see the end of the last line.

Is there a function in ggplot which I can add below the last line, which will be read, but will do nothing, preventing me from having to deal with the + signs?

For example, Python has the pass statement, which acts as a filler when writing code. Does something similar exist in ggplot?

Upvotes: 3

Views: 350

Answers (3)

wjchulme
wjchulme

Reputation: 2028

A perhaps safer option is to use NULL. This does literally nothing whereas geom_blank() adds a layer. The use of NULL after a ggplot chain is a deliberate design feature.


library('ggplot2')

plot0 = ggplot(iris) +
  geom_point(aes(x=Sepal.Width, y=Sepal.Length, colour=Species))

plot1 = ggplot(iris) +
  geom_point(aes(x=Sepal.Width, y=Sepal.Length, colour=Species))+
#  scale_y_continuous()+
  NULL

plot2 = 
  ggplot(iris) +
  geom_point(aes(x=Sepal.Width, y=Sepal.Length, colour=Species))+
#  scale_y_continuous()+
  geom_blank()


length(plot0$layers)
#> [1] 1

length(plot1$layers)
#> [1] 1

length(plot2$layers)
#> [1] 2

object.size(plot0) < object.size(plot1)
#> [1] FALSE

object.size(plot0) < object.size(plot2)
#> [1] TRUE

Upvotes: 1

user9817463
user9817463

Reputation: 11

ggplot() + theme()

or 

ggplot() + list()

or 

ggplot() + invisible(list())

do nothing and do not add a layer. The last one also won't print anything if executed on its own.

Upvotes: 1

Technophobe01
Technophobe01

Reputation: 8676

You have two options

a. geom_blank()

b. gp <- gp +.

The gp <- gp + solution is my personal recommendation. It allows for simple addition and removal of ggplot components and as a byproduct gives you a gp gglpot structure you can interogate via RStudio:

gp <- ggplot()
gp <- gp + geom_col() 
gp <- gp + geom_line() 
# gp <- gp + scale_y_continuous()
gp

this achieves the same result as the end of line + without the extra superfluous function call. As an alternative, you might consider using geom_blank() if you have a large code base to update then the latter geom_blank() may be an easier choice. (This assumes you still debugging or tweaking your plots)

ggplot() +
geom_col() +
geom_line() +
# scale_y_continuous() +
geom_blank()

my sense is this is a question of aesthetics. I am not a fan of trailing + because of the potential for drop through to other code (line overrun). Such situations can create interesting and awkward debugging situations.

Upvotes: 0

Related Questions