I'm using a script in RStudio which produces several graphs. This is a (very simplified) example of how it looks
mpg = mtcars$mpg
cyl = mtcars$disp
hp = mtcars $hp
plot(mpg, cyl)
dev.new()
plot(mpg,hp)
dev.new()
plot(hp, cyl)
When I run this in RStudio (using the "Source" key), the first graph stays in the "Plots" viewer in the lower right of the Rstudio window (I would actually like it to be a separate window as well if anyone knows how to do that), and the second and third graphs are produced as individual windows separate from the RStudio screen.
When I try to change the size of these windows (by pressing the full screen button in the top right corner, or by dragging the side), the most recently generated window is fine and resizes, but the previous one turns grey and the graph stops showing.
I thought it might have something to do with the fact that the graphics devices are inactive, but the most recently generated one, which rescales fine, also says it's inactive.
Any help would be appreciated! Thanks.
Edit: here is the output of sessionInfo():
> sessionInfo()
R version 4.0.0 (2020-04-24)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)
Matrix products: default
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] DescTools_0.99.36 dplR_1.7.1
loaded via a namespace (and not attached):
[1] Rcpp_1.0.4.6 mvtnorm_1.1-1 lattice_0.20-41 matrixStats_0.56.0 XML_3.99-0.3 png_0.1-7
[7] digest_0.6.25 MASS_7.3-51.5 R.methodsS3_1.8.0 grid_4.0.0 plyr_1.8.6 signal_0.7-6
[13] magrittr_1.5 stringi_1.4.6 rstudioapi_0.11 R.oo_1.23.0 R.utils_2.9.2 Matrix_1.2-18
[19] boot_1.3-24 tools_4.0.0 stringr_1.4.0 compiler_4.0.0 expm_0.999-4
It's not a perfect answer, but here's what I found to work around it:
When you generate a number of graphics devices the most recently generated one remains "active" while the rest are inactive. If you don't close the active window and resize the others, it works fine. But if you close the active window and then try to resize the others, they just go grey. Somehow, closing the active window shuts down something and stops you from being able to resize the graphs at all. Weird.
Related
I was interested in converting my numericInputs into autonumericInputs from the shinywidgets package. My primary issue is that even when the default value of autonumericInput is NA/NULL/"", the output still reads as 0, and counts it as having 1 character. From a server side, I'd like to say one of the following: if(is.null(x) | is.na(x) | x == "" | nchar(x) == 0), then another object will not run.
For some numericInputs this isn't an issue because I can just say it needs to be greater than 0. However a 0 is acceptable for some inputs, and I want to force the input to be filled in. I suppose I could default the input to "-1" and then say greater than or equal to 0, requiring at least some interaction with the input, but that seems messy to me.
This is an example with regular numericInput on the first column, and autonumericInput on the second column. The text below each input should not show if the field is blank.
library(shiny)
library(shinyWidgets)
ui <- fluidPage(
column(width = 2,
numericInput("Test", "Numeric Input Test", value = NULL),
textOutput("Test2")
),
column(width = 2,
autonumericInput("Test1", "AutoNumeric Input Test", value = NULL),
textOutput("Test3")
)
)
server <- function(input, output, session) {
output$Test2<-renderText({
req(input$Test, nchar(input$Test) > 0, !is.na(input$Test), !is.null(input$Test))
paste0("Numeric Input reads as: ",input$Test,". Number of characters: ",nchar(input$Test))
})
output$Test3<-renderText({
req(input$Test1, nchar(input$Test1) > 0, !is.na(input$Test1), !is.null(input$Test1))
paste0("Autonumeric Input reads as: ",input$Test1,". Number of characters: ",nchar(input$Test1))
})
}
shinyApp(ui, server)
I have not found a way to handle these, and I was hoping somebody had an idea. Perhaps this just needs to be a feature request or issue on the package github: https://github.com/dreamRs/shinyWidgets/ I've read through the documentation and at least I'm not seeing any functions that I think would solve it: https://search.r-project.org/CRAN/refmans/shinyWidgets/html/autonumericInput.html
Why I'd like to switch to autonumericInput is because I like the greater control, simplifying user inputs. Limiting accepted values, always showing decimal places, limited points accepted after a decimal, showing symbols like %.
Perhaps this is an out of date package problem? I'm hoping that's not the case, else I need to update packages of several machines some of which don't have internet access, here is my sessionInfo():
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)
Matrix products: default
locale:
[1] LC_COLLATE=English_United States.1252
[2] LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] shinyWidgets_0.6.0 shiny_1.6.0
loaded via a namespace (and not attached):
[1] Rcpp_1.0.2 withr_2.4.2 digest_0.6.27
[4] later_1.2.0 mime_0.7 R6_2.4.0
[7] jsonlite_1.6 lifecycle_1.0.0 xtable_1.8-4
[10] magrittr_1.5 cli_2.5.0 cachem_1.0.4
[13] rlang_0.4.11 promises_1.2.0.1 jquerylib_0.1.4
[16] bslib_0.2.4 ellipsis_0.3.2 tools_3.6.1
[19] httpuv_1.6.1 fastmap_1.1.0 compiler_3.6.1
[22] sessioninfo_1.1.1 htmltools_0.5.1.1 sass_0.3.1
I don't know if this is related, but I think shinywidgets uses javascript, and this stack listing mentions javascript reading empty lines as 0: Node.js + Lazy: how to prevent an empty line being read as "0"? All else fails, I'll just stick with regular numericInput. A pre-emptive thanks for anyone reading this!
I am using the following chunk in an Rnw file.
<<scatter,dev='pdf',include=TRUE>>=
a <- rnorm(1e4)
b <- rnorm(1e4)
plot(a,b)
#
I knitted the same Rnw file under R3.6.3 and R4.0.2, respectively. However, the PDF figure generated under R4.0.2 is much larger (~8 times) than the one from R3.6.3.
I am wondering if there is a way (eg. a chunk option) to reduce the PDF size under R4.0.2.
The session information of two different versions of R is shown below.
R version 4.0.2 (2020-06-22)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 16299)
Matrix products: default
locale:
[1] LC_COLLATE=English_Australia.1252 LC_CTYPE=English_Australia.1252
[3] LC_MONETARY=English_Australia.1252 LC_NUMERIC=C
[5] LC_TIME=English_Australia.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] knitr_1.29
loaded via a namespace (and not attached):
[1] compiler_4.0.2 magrittr_1.5 tools_4.0.2 stringi_1.4.6 highr_0.8 stringr_1.4.0
[7] xfun_0.16 evaluate_0.14
R version 3.6.3 (2020-02-29)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 16299)
Matrix products: default
Random number generation:
RNG: Mersenne-Twister
Normal: Inversion
Sample: Rounding
locale:
[1] LC_COLLATE=English_Australia.1252 LC_CTYPE=English_Australia.1252
[3] LC_MONETARY=English_Australia.1252 LC_NUMERIC=C
[5] LC_TIME=English_Australia.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] knitr_1.28
loaded via a namespace (and not attached):
[1] compiler_3.6.3 magrittr_1.5 tools_3.6.3 stringi_1.4.6 highr_0.8 stringr_1.4.0
[7] xfun_0.12 evaluate_0.14
The only reference to pdf in the release notes lately has been in relation to the dingbats font. Looking at the reference, it says (emphasis mine):
useDingbats logical. Should small circles be rendered via the
Dingbats font? Defaults to TRUE, which produces smaller and better
output. Setting this to FALSE can work around font display problems in
broken PDF viewers: although this font is one of the 14 guaranteed to
be available in all PDF viewers, that guarantee is not always
honoured.
You can change the default via
pdf.options(useDingbats = TRUE)
in the first code chunk of your Rnw document. This change is global to your document. If you'd rather change it only for a specific code chunk, you may use the chunk option dev.args, e.g.,
<<scatter, dev='pdf', dev.args=list(useDingbats = TRUE)>>=
a <- rnorm(1e4)
b <- rnorm(1e4)
plot(a,b)
#
When using R Studio, I usually just work with an .R file stacked on top of the Console. I keep the other panes (Environment, History, Files, etc) hidden.
But whenever I plot a graph, the other panes automatically pop out of the side bar to show me the Plot pane. Because I work on a laptop, this makes everything too small to see. By clicking the Zoom button on the Plots pane, I can get the plot also show up in a new window, but does not prevent the Plots pane from showing up.
Is there a way to "disable" the Plots pane in R Studio, and force plots show up in a new window?
> sessionInfo()
R version 3.2.3 (2015-12-10)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 7 x64 (build 7601) Service Pack 1
locale:
[1] LC_COLLATE=English_United States.1252
[2] LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils
[5] datasets methods base
other attached packages:
[1] ggplot2_2.2.1 jsonlite_1.4
[3] data.table_1.10.4
loaded via a namespace (and not attached):
[1] labeling_0.3 colorspace_1.2-6
[3] scales_0.4.1 lazyeval_0.2.0
[5] plyr_1.8.4 tools_3.2.3
[7] gtable_0.1.2 tibble_1.3.0
[9] curl_2.5 Rcpp_0.12.10
[11] grid_3.2.3 munsell_0.4.2
>
The dev.new() function will open a new plot window, which then becomes the target for all plots.
If you wish to open another window you can run the command a second time to open a second window.
dev.off() will shut down the window (in the order they were opened by default).
You can see how to control multiple graphics devices in the documentation here.
In RStudio, the default graphics device is normally "RStudioGD". You can change that to something else: the normal choices are "windows" on Windows, "quartz" on MacOS, "X11" on Linux. So for example, use
options(device = "quartz")
in your RStudio session on a Mac and you'll get the regular MacOS graphics window.
Try using the windows command before your plot call.
windows();(mpg ~ wt, mtcars)
The plot should pop-up in its own window whilst the pane stays minimized.
Commenting the following lines in C:\Program Files\RStudio\R\Tools.R seems to work (it may be necessary to edit the file as administrator):
# set our graphics device as the default and cause it to be created/set
.rs.addFunction( "initGraphicsDevice", function()
{
# options(device="RStudioGD")
# grDevices::deviceIsInteractive("RStudioGD")
grDevices::deviceIsInteractive()
})
If you want all plots in the current script to appear on a separate window, this should do the job:
dev.new(noRStudioGD = TRUE)
*Tested on RStudio version 1.4.1106 for Windows with R version 4.0.5 (2021-03-31)
For a 'permanent' solution, the answer by Rubén Fernández-Casal should work well.
You can force RStudio to show plots in the Source window if you use R Markdown. In a Rmd file, plots are shown together with code; it's called an R Markdown notebook. You can set the size of the plots too, in what is called an R code chunk:
```{r fig.height = 2, fig.width = 3}
plot(mpg ~ wt, mtcars)
```
When you run the chunk, the plot is shown below it.
If you want to set the plot size for the whole notebook, set the package option using opts_knit and opts_chunk, for example:
```{r setup}
library(knitr)
opts_knit$set(global.par = TRUE)
opts_chunk$set(fig.width = 4.5, fig.height = 3.5)
```
For more information, see here and here.
I cannot provide a reproducible example for this, but I wanted to ask about a curious behavior of the ggsave() function: I am experiencing the following:
I create a ggplot, by assigning a plot to a variable, let's say p
I evaluate p and - as it should be - the plot is displayed in the plot pane of RStudio.
My next call is ggsave(plot = p, filename = "plot.pdf"), the plot is
written to plot.pdf and R gives me the dimensions of the written plot on the console.
The strange thing is: When the RStudio plot pane is active (i.e. a RStudio graphics device is open), the dimensions of the plot written by ggsave() in step 3 are different than when no RStudio graphics device is open. Consequently, the scalings of the written plot are off. In other words: Step 3 produces different results depending on the execution of step 2. Does anyone know why this is the case? I thought that ggsave() is totally independent of any RStudio device. Or does this only happen on my machine?
This is my sessionInfo():
R version 3.4.1 (2017-06-30)
Platform: x86_64-apple-darwin15.6.0 (64-bit)
Running under: macOS Sierra 10.12.6
Matrix products: default
BLAS: /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/3.4/Resources/lib/libRlapack.dylib
locale:
[1] de_DE.UTF-8/de_DE.UTF-8/de_DE.UTF-8/C/de_DE.UTF-8/de_DE.UTF-8
attached base packages:
[1] stats graphics grDevices utils datasets methods base
other attached packages:
[1] ggplot2_2.2.1 scales_0.4.1 data.table_1.10.4
loaded via a namespace (and not attached):
[1] ggrepel_0.6.12 Rcpp_0.12.12 digest_0.6.12 grid_3.4.1 plyr_1.8.4 gtable_0.2.0 rlang_0.1.2 lazyeval_0.2.0
[9] labeling_0.3 RColorBrewer_1.1-2 tools_3.4.1 munsell_0.4.3 compiler_3.4.1 colorspace_1.3-2 sciplot_1.1-1 tibble_1.3.3
Expanding on Andrey Kolyadin's comment above...
From ggsave() documentation 'width, height Plot size in units ("in",
"cm", or "mm"). If not supplied, uses the size of current graphics
device.'
If we dig into the code behind ggsave, there's a snippet that says:
if (any(is.na(dim))) {
if (length(grDevices::dev.list()) == 0) {
default_dim <- c(7, 7)
}
else {
default_dim <- grDevices::dev.size() * scale
}
...
Which means that if the dimensions are not specified by the user:
if there's no active graphics device at all, default dimensions are 7 x 7;
if there's at least one active graphics device, use the dimensions of the last one (multiplied by scale, which defaults to 1).
RStudio's plot pane is a graphics device. So ggsave's behaviour differs depending on whether there's anything there.
I am able to create a dynamic 3D scatter plot using rgl and knitr, but the legend created by the legend3d is missing. I guess it may be related to the implementation of hook_webgl. Any idea? Thank you.
Here are the code I used to create the 3D scatter plot:
```{r fig1, webgl=TRUE, fig.width=10 , fig.height = 10}
library(knitr)
library(rgl)
knit_hooks$set(webgl = hook_webgl)
x <- sort(rnorm(1000))
y <- rnorm(1000)
z <- rnorm(1000) + atan2(x,y)
plot3d(x, y, z, col=rep(1:2,each=500))
legend3d("topright",legend=c("Type 1","Type 2"), pch=16, col=1:2)
```
My session info:
R version 3.1.3 (2015-03-09)
Platform: i386-w64-mingw32/i386 (32-bit)
Running under: Windows Server 2008 R2 x64 (build 7601) Service Pack 1
locale:
[1] LC_COLLATE=English_United States.1252
[2] LC_CTYPE=English_United States.1252
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] stats graphics grDevices utils datasets methods
[7] base
other attached packages:
[1] knitr_1.11 rgl_0.95.1441 dplyr_0.4.3
loaded via a namespace (and not attached):
[1] assertthat_0.1 DBI_0.3.1 digest_0.6.8 evaluate_0.8
[5] formatR_1.2.1 highr_0.5.1 htmltools_0.3 lazyeval_0.1.10
[9] magrittr_1.5 parallel_3.1.3 R6_2.1.2 Rcpp_0.12.3
[13] rmarkdown_0.9.5 stringi_0.5-5 stringr_1.0.0 tools_3.1.3
[17] yaml_2.1.13
Looks like a bug in the rglwidget code; I'll look into it.
I recommend a few changes to your programming style. These won't help with this bug, but may save you problems later:
Use the setupKnitr() function rather than setting the knitr hook yourself. The link to knitr is likely to change, but setupKnitr() will protect you from problems.
Rather than saying webgl=TRUE in the chunk options, make a call to the rglwidget::rglwidget() function when you want to draw the plot. I'm hoping to make this call unnecessary in the future (more like what happens with regular plots), but for now it's the better tested way to do things.
Update your R. Current version is 3.2.3, with 3.2.4 due quite soon.
Use rgl and rglwidget from R-forge. They are newer than the CRAN versions, and have several bug fixes. Hopefully soon they'll have the legend bug fix too.
Added later:
This wasn't so much a bug as something that wasn't implemented. Now it is, if you use rgl version 0.95.1456 and rglwidget version 0.1.1456, both available from R-forge. Not much testing yet, so there may be more changes over the next while.