Persistent invalid graphics state error when using ggplot2

135
votes

I believe my dataframe is okay and my code is okay. In fact, I have eliminated parts of the dataframe and most of the graphing code to make things as basic as possible. But still, I get:

Error in .Call.graphics(C_palette2, .Call(C_palette2, NULL)) : 
  invalid graphics state

What is wrong here? Here is the data: 

 date   trt var val
1/8/2008    cc  sw5 0.2684138
1/8/2008    cc  sw15    0.2897586
1/8/2008    cc  sw5 0.2822414
2/8/2008    cc  sw5 0.2494583
2/8/2008    cc  sw5 0.2692917
2/8/2008    cc  sw15    0.2619167
2/8/2008    cc  sw5 0.204375
3/8/2008    cc  sw5 0.2430625
3/8/2008    cc  sw5 0.2654375
3/8/2008    cc  sw5 0.2509583
3/8/2008    cc  sw5 0.2055625
1/8/2008    ccw sw15    0.2212414
1/8/2008    ccw sw5 0.3613448
1/8/2008    ccw sw5 0.2607586
2/8/2008    ccw sw5 0.2087917
2/8/2008    ccw sw15    0.3390417
2/8/2008    ccw sw5 0.2436458
2/8/2008    ccw sw5 0.290875
3/8/2008    ccw sw5 0.20175
3/8/2008    ccw sw15    0.328875
3/8/2008    ccw sw5 0.2328958
3/8/2008    ccw sw5 0.2868958

When I work with this data, I specify dates like this:

df<-df[order(as.Date(df$date,format="%d/%m/%Y")),,drop=FALSE]

and here I want to make a scatterplot:

ggplot(data = df,aes(x = date,y = val)) + 
 geom_point(aes(group = trt))
7
rggplot2
Have you tried (re)installing ggplot2? – Paul Hiemstra
I added this as an answer, in that way other people can find the answer more easily. You can accept the answer by clicking on the tick mark right next to my answer. Just to confirm, you had ggplot2 installed originally? – Paul Hiemstra
Yes, it was already installed. – Nazer
restarting R session solved my problem. – Chris
FYI - this issue can also arise if you attempt to construct use figure dimensions that are too small (e.g. attempting to provide dimensions in inches without specifying the unit). – Keith Hughitt

7 Answers

402
votes

I ran into this same error and solved it by running: 

dev.off()

and then running the plot again. I think the graphics device was messed up earlier somehow by exporting some graphics and it didn't get reset. This worked for me and it's simpler than reinstalling ggplot2.

11
votes

I found this to occur when you mix ggplot charts with plot charts in the same session. Using the 'dev.off' solution suggested by Paul solves the issue.

9
votes

The solution is to simply reinstall ggplot2. Maybe there is an incompatibility between the R version you are using, and your installed version of ggplot2. Alternatively, something might have gone wrong while installing ggplot2 earlier, causing the issue you see.

4
votes

You likely don't need to reinstall ggplot2

Solution: go back to plot that didn't work previously. Take the below console output for example. The figure margins (the window that displays your plots) were too small to display the pairs(MinusInner) plot. Then when I tried to make the next qplot, R was still hung up on previous error.

pairs(MinusInner) Error in plot.new() : figure margins too large qplot(Sample.Type, BAE,data=MinusInner, geom="boxplot") Error in .Call.graphics(C_palette2, .Call(C_palette2, NULL)) : invalid graphics state

I fixed the first error by expanding the plot window and rerunning the pairs(MinusInner) plot. Then blam, it worked.

pairs(MinusInner) qplot(Sample.Type, BAE,data=MinusInner, geom="boxplot")

0
votes

I solved this by clearing all the plots in the console and then making sure the plot area was large enough to accommodate what I was creating.

0
votes

try to get out grafics with x11() or win.graph() and solve this trouble.

0
votes

Restarting R (Menu>Sessions>Restart R or ⌘⇧F10) and re-running my code fixed the problem.