Resize Points in Existing Plot Fails

[Dario Strbenac]

If I have a function which does lots of processing and returns a ggplot object, I don't want to rerun the function and do all of the time-consuming calculations again, but it seems impossible to change the point size of plot points. The object is also huge when saved to disk.

> class(myPlot)
[1] "gg"     "ggplot"

myPlot <- myPlot + geom_point(size = 1) # Doesn't change appearance.
myPlot
myPlot <- myPlot + geom_point(size = 10) # Doesn't change appearance.
myPlot

> format(object.size(myPlot), units = "KB")
[1] "26 Kb"

> save(myPlot, file = "myPlot.obj")     # Why 198 MB on disk ?

#############
## $ ls -l myPlot.obj
## -rw-r----- 1 dario stgrad 208140875 Mar 20 16:52 myPlot.obj
#############

> sessionInfo()
R version 3.1.2 (2014-10-31)
Platform: x86_64-pc-linux-gnu (64-bit)

locale:
[1] C

attached base packages:
[1] parallel  stats     graphics  grDevices utils     datasets  methods   base    

other attached packages:
[1] ggplot2_1.0.0       ClassifyR_1.1.32    BiocParallel_1.0.3  Biobase_2.26.0      BiocGenerics_0.12.1 limma_3.22.4

[Doug Mitarotonda]

Well, because you didn’t provide a reproducible example, it is hard to help you. When I make a small example on my own what you suggest works fine. 

> library(ggplot2)
> dft <- data.frame(A=1:3, B=4:6)
> ggplot(dft, aes(A,B)) + geom_point()
> p <- ggplot(dft, aes(A,B)) + geom_point()
> p + geom_point(size=1) # small points

> p + geom_point(size=10) # large points

--
--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility
 
To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

---
You received this message because you are subscribed to the Google Groups "ggplot2" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ggplot2+u...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Dario Strbenac]

Hello,

That simple example works for me, too. I have a reproducible example which does not work. You many copy and paste it directly into R. It fails if you resize the points more than once.

library(ggplot2)

aData <- data.frame(A = factor(sample(c('A', 'B'), 50, replace = TRUE)),
                    B = factor(sample(c('C', 'D'), 50, replace = TRUE)),
                    C = factor(rep(seq(10, 100, 10), 5)),
                    D = rnorm(50, 10, 10))

overlapPlot <- ggplot(aData, aes(x = C, y = D, colour = A, shape = B)) + geom_line(size = 2) + geom_point(size = 3)

overlapPlot <- overlapPlot + geom_point(size = 10)
overlapPlot     # This works.
overlapPlot <- overlapPlot + geom_point(size = 1)
overlapPlot     # This fails.

> sessionInfo()
R version 3.1.2 (2014-10-31)

Platform: i386-w64-mingw32/i386 (32-bit)

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252    LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                           LC_TIME=English_United States.1252   

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base    

other attached packages:
[1] ggplot2_1.0.0

loaded via a namespace (and not attached):
 [1] colorspace_1.2-4 digest_0.6.8     grid_3.1.2       gtable_0.1.2     labeling_0.3     MASS_7.3-39      munsell_0.4.2  
 [8] plyr_1.8.1       proto_0.3-10     Rcpp_0.11.4      reshape2_1.4.1   scales_0.2.4     stringr_0.6.2

[Dennis Murphy]

The plot does not "fail" - the way you've constructed it makes it
impossible to see the result of the last call because the fill color
of points is the same in all three calls. Here's a modified version of
your code; the line is removed to better see what is going on. The
point shape is also changed to unfilled to show that the last call
indeed works.

ggplot(aData, aes(x = C, y = D, colour = A)) +
geom_point(size = 4, shape = 21)

last_plot() + geom_point(size = 10, shape = 21)
last_plot() + geom_point(size = 2, shape = 21)


As for your original plot, one suggestion is to change to an unfilled
point shape (21-25) as follows:

overlapPlot <- ggplot(aData, aes(x = C, y = D, colour = A, shape = B)) +

geom_line(size = 2) + geom_point(size = 3) +
scale_shape_manual(values = c(21, 22))

last_plot() + geom_point(size = 10)
last_plot() + geom_point(size = 1)


ggplot2 is based on grid graphics, and grid operates on the "painter's
model" - to quote Paul Murrell in Chapter 5 of R Graphics, 2e:

"Like the traditional system, all grid output occurs on the current
device, and later output obscures any earlier output that it overlaps
(i.e.,output follows the “painters model”)."

Dennis

[Dario Strbenac]

If it's like a painter's canvas, why does using plot.new() before the last plot not work as desired ?

[Dennis Murphy]

Because plot.new() is a base graphics function, not grid. You start a
new graphics page in grid with grid.newpage(). Calling grid.newpage()
wipes out the previous contents; that's its purpose. Since I have no
idea what "work as desired" means, I can't attempt to help further.

Dennis

On Sat, Mar 21, 2015 at 12:00 AM, Dario Strbenac <dario....@gmail.com> wrote:
> If it's like a painter's canvas, why does using plot.new() before the last
> plot not work as desired ?
>

[Dario Strbenac]

What I mean is that I would expect the following code to clear the canvas, enabling the size 1 points to be visible.

overlapPlot <- overlapPlot + geom_point(size = 10)
overlapPlot

grid.newpage()