R : Error looping in ggplot2

Question

I'm trying to use a loop to simplify a ggplotcommand, but it only assigns the last part of the loop to my ggplot object.

plot.test = ggplot(data = plot.df, aes(x = xvariable)) + 
  geom_line(aes(y = yvariable)) 

for (i in 1:6){
  plot.test = plot.test + geom_point(aes(y = ypoints[i], x = xpoints[i])) 
}

So, the problem is, that after running the loop, plot.test does only contain the original line and the last point from the loop, i.e. point (x[6],y[6])

Any ideas on how to solve this?

Beasterfield · Accepted Answer · 2014-04-02 13:49:48Z

3

Why would you want to do that?

Is

ggplot( data = plot.df, aes( x = xvariable, y = yvariable ) ) +
  geom_line() +
  geom_point( data = plot.df[1:6,], aes( y = ypoints, x = xpoints ) )

what you are looking for?

Edit:

In my original answer I wrote that You cannot add multiple geometric objects of the same type to one ggplot. This is obviously wrong as the following example states:

dt <- data.frame( x = rnorm(1:10), y = rnorm(1:10) )
ggplot( mapping = aes( x = x, y = y ) ) +
  geom_point( data = dt[1:5,] ) +
  geom_point( data = dt[6:10,] )

However I have really difficulties to think of an example where this would be necessary. From my experience I would always presume that the reason for using two times the same geom in the same plot is a result of a bad data model.

edited Apr 2, 2014 at 13:49

answered Apr 2, 2014 at 13:18

Beasterfield

7,1332 gold badges43 silver badges49 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

pangia · Accepted Answer · 2014-04-02 13:30:29Z

1

Here's a dataset to work with.

plot.df <- data.frame(xvariable=1:6, yvariable=1:6+rnorm(6,0,0.3), xpoints=1:6+rnorm(6,0,0.1), ypoints=1:6+rnorm(6,0,0.1))

It IS possible to add multiple geometric objects of the same type to one ggplot. For example:

ggplot(data = plot.df, aes(x = xvariable)) + 
  geom_line(aes(y = yvariable)) +
  geom_point(aes(y=ypoints[1], x=xpoints[1])) +
  geom_point(aes(y=ypoints[2], x=xpoints[2])) +
  geom_point(aes(y=ypoints[3], x=xpoints[3])) +
  geom_point(aes(y=ypoints[4], x=xpoints[4])) +
  geom_point(aes(y=ypoints[5], x=xpoints[5])) +
  geom_point(aes(y=ypoints[6], x=xpoints[6]))

But that's a pain, and you're rightly seeking to avoid that sort of code replication. The answer depends on what you're looking for. Different formatting for each point? Try this:

ggplot(data = plot.df, aes(x = xvariable)) + 
  geom_line(aes(y = yvariable)) +
  geom_point(data=transform(plot.df, pointsID=as.factor(1:6)), 
             aes(y=ypoints, x=xpoints, color=pointsID))

You can alter the colors (or choices for other formatting parameters) through calls to scale_colour_discrete and related ggplot components.

PS - Your original strategy failed because ggplot waits until print time to evaluate arguments such as the i in xpoints[i]; even though you see 1 point, you're actually plotting 6 of the same point right on top of one another.

answered Apr 2, 2014 at 13:30

pangia

1,16410 silver badges11 bronze badges

1 Comment

blah_crusader Over a year ago

That's exactly what I was trying to do. I restructured my data already by realizing geom_point can take on whole dataframes of x,y coordinates based on the answer above. But checked this , and works as well.

Thomas Sibley · Accepted Answer · 2014-06-16 20:53:18Z

While reworking the data structure to avoid loops like this is the better answer, sometimes you really need to evaluate the expression immediately instead of closing over a variable and evaluating the expression later. You can use ggplot2's aes_string() to do just that:

plot.test = ggplot(data = plot.df, aes(x = xvariable)) + 
  geom_line(aes(y = yvariable)) 

for (i in 1:6){
  plot.test = plot.test + geom_point(aes_string(y = ypoints[i], x = xpoints[i])) 
}

In this case you're effectively hardcoding the point values. That is, it's equivalent to:

  plot.test = plot.test + geom_point(aes(y = 2, x = 3))

You can also hardcode only the index, and still dynamically refer to whatever is in ypoints and xpoints at a later time:

  plot.test = plot.test + geom_point(aes_string(y = paste0("ypoints[", i, "]"), x = paste0("xpoints[", i, "]")))

This is ugly and unwieldy and a bit of a hack, but it works and is sometimes just what you need if you can't liberally reorganize the data structure for whatever reason.

Collectives™ on Stack Overflow

R : Error looping in ggplot2

3 Answers 3

Comments

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related