I seem to have discovered an odd behaviour with the knit Word command in RStudio
This works:
```{r qplot, fig.width = 6, fig.height=6, message=FALSE}
library(ggplot2)
summary(cars)
qplot(speed, dist, data = cars) + geom_smooth()
````
this does not work
```{r q plot, fig.width = 6, fig.height=6, message=FALSE}
library(ggplot2)
summary(cars)
qplot(speed, dist, data = cars) + geom_smooth()
```
returning this message:
pandoc.exe: Could not find image `./test_files/figure-docx/q%20plot.png', skipping...
The issue seems to be with the name of the chunk (i.e. qplot vs. q plot). When there is a space in the chunk name the plot does not render.
It only seems to affect the rendering of Word documents. Rendering html works fine.
I'm using RStudio 0.98.1028 and R3.1.1 on windows 7.
Has anyone else encountered this behaviour?
update
a space after the chunk name also seems to elicit the same behaviour:
this does not work
```{r q_plot , fig.width = 6, fig.height=6, message=FALSE}
library(ggplot2)
summary(cars)
qplot(speed, dist, data = cars) + geom_smooth()
```
Posting the solution in case someone runs across this in future.
From Ben Bolker in the comments Avoid spaces and periods . in chunk labels and directory names as stated in the knitr documentation http://yihui.name/knitr/options.
This error only seems to affect making plots using knitWord. Code chunks with labels that contain spaces and that don't have plotting commands render normally. knitHTML also seems to work fine regardless of if chunk labels have a space or not.
# Let's make a plot
```{r ugly plot}
plot(btc_prices)
```
should apparently be
# Let's make a plot
```{r ugly_plot}
plot(btc_prices)
```
So no spaces... otherwise you'll waste hours googling and crying.
Related
I am transitioning to Bookdown from Markdown. Although it seems as though this should be straightforward I am getting an unexpected knit failure once I have added the 3rd (usually) code chunk in a chapter (not index.rmd) file. I have tried using code from the example code by Allaire and Xie with the same result. The problem does not occur if all the markdown code is in one file i.e. no chapters. Here is the sample code from the index and 1st chapter files. Sorry about the formatting; I'm not sure how to format markdown code in stackoverflow so the 3 back tics consistently show.
---
classoption: openany # Removes blank pages (Arrgh!!)
site: "bookdown::bookdown_site"
output:
bookdown::tufte_book2:
latex_engine: xelatex
---
```{r setup, include=FALSE}
library(ggplot2)
# invalidate cache when the tufte version changes
knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion("tufte"))
```
And this is the first chapter code:
# Figures
## Margin Figures
Images and graphics play an integral role in Tufte's work. To place figures in the margin you can use the **knitr** chunk option `fig.margin = TRUE`. For example:
```{r fig-margin, fig.margin = TRUE, fig.cap = "MPG vs horsepower, colored by transmission.", fig.width=3.5, fig.height=3.5, cache=TRUE, message=FALSE}
library(ggplot2)
mtcars2 <- mtcars
mtcars2$am <- factor(
mtcars$am, labels = c('automatic', 'manual')
)
ggplot(mtcars2, aes(hp, mpg, color = am)) +
geom_point() + geom_smooth() +
theme(legend.position = 'bottom')
```
In fact, you can include anything in the margin using the **knitr** engine named `marginfigure`. Unlike R code chunks ```` ```{r} ````, you write a chunk starting with ```` ```{marginfigure} ```` instead, then put the content in the chunk. See an example on the right about the first fundamental theorem of calculus.
```{marginfigure}
We know from _the first fundamental theorem of calculus_ that for $x$ in $[a, b]$:
$$\frac{d}{dx}\left( \int_{a}^{x} f(u)\,du\right)=f(x).$$
```
It fails here
knitr::kable(
mtcars[1:6, 1:6], caption = 'A subset of mtcars.'
)
Deleting the code after "It fails here" makes it knit properly for me.
Dropbox link with scripts, log, and pdf's.
I am producing a rmarkdown document, knitting to PDF and have a figure (figure 1) and a table (table 1) where the table explains the figure in more detail. I have no issue giving either of them a standard caption but I would like to change the table caption to be "Explanation of Figure 1". Is there any way of doing this?
The code chunks are listed below, please let me know if I need to provide more information:
YAML:
- \usepackage{caption} #and several others
output:
bookdown::pdf_document2:
keep_tex: no
latex_engine: xelatex
Code Chunks:
Figure 1:
```{r figure-1, fig.cap="Figure"}
ggplot()
```
Table 1:
```{r table, fig.cap="Explanation of Figure \#ref(fig:figure-1)"}
knitr
kableExtra::kable(caption = "Explanation of Figure \#ref(fig:figure-1)")
```
The main error message with one backslash is "Error: '#' is an unrecognized escape in character string" and suggests I forgot to quote character options, which is not true.
With two backslashes the document knits but produces the caption "Explanation of Figure reffig:table"
3 backslashes: the same error as with 1.
4 backslashes: the error is "pandoc-citeproc: reference ref not found. ! Package caption Error: \caption outside float."
Appreciate any suggestions!
Just a workaround, but may helps.
The \\#ref(fig:xxx) option works well when knitting to a html_document2.
To me pdf - creation worked fine when using pandoc in the terminal.
E.g.:
---
title: "Cross ref"
output:
bookdown::html_document2:
collapsed: no
theme: readable
toc: yes
link-citations: yes
---
```{r firstplot, fig.cap = "A plot with points." }
library(ggplot2)
plot_A = ggplot(data = data.frame(x = c(1:10),
y = seq(3, 8, length.out = 10)),
aes(x = x, y =y))+
geom_point()
plot_A
```
Now a second plot with a reference to Fig.: \#ref(fig:firstplot).
```{r secondplot, fig.cap = "This is the same as Fig.: \\#ref(fig:firstplot)
but now with a red line." }
library(ggplot2)
plot_A + geom_line(alpha = .75,col = "red")
```
after knitting just move to the folder containing the html and using pandoc
pandoc mini_ex-crossref.html -o mini_ex.pdf
I tried many different approaches text references, chunk captions, caption argument in the kable function and I´m sure there is a clever solution somewhere, so here is just a workaround with pure Latex.
Add a latex chunk with a label before the chunk with the figure:
```{=latex}
\begin{figure}
\caption{Figure 1}
\label{Fig-1}
```
```{r figure-1, echo = FALSE}
ggplot(mtcars) +
geom_point(aes(cyl, gear))
```
```{=latex}
\end{figure}
```
Now you can refer to Fig-1 in your latex-caption for the table with normal latex code \ref{Fig-1}:
```{=latex}
\begin{table}
\caption{Explanation of Figure \ref{Fig-1}}
```
```{r table}
kableExtra::kable(x = mtcars)
```
```{=latex}
\end{table}
```
Notes:
* In my opinion this is just a workaround.
* It´s not possible to use the chunk option fig.cap = "" and the latex code in parallel
J_F referenced Yihui Xie's excellent explanation of using text references in RMarkdown (https://bookdown.org/yihui/bookdown/markdown-extensions-by-bookdown.html#text-references), which you can use for figure and table captions that require more complicated things than plain text (e.g., formatting, cross-references, etc.). This may be a more flexible solution overall than remembering to escape the backslash in Robert's answer, and does not require a workaround with LaTeX.
As Yihui explains, all you need to do is define a text reference on a single line in markdown and reference that in the chunk option "fig.cap" or the "caption" parameter in knitr::kable(). Just be careful to make sure that each text reference is one paragraph that does not end in a white space.
Here's a basic example.
---
title: "Cross-referencing figures and tables within captions."
output: bookdown::pdf_document2
editor_options:
chunk_output_type: console
---
```{r load-packages}
library(knitr)
library(flextable)
```
(ref:first-fig-caption) Here's a complicated figure caption for the first figure, which can include complicated text styling like $m^2$ and references to other elements in the document, like Table \#ref(tab:mtcars) or Fig. \#ref(fig:cars).
```{r pressure, fig.cap = '(ref:first-fig-caption)'}
plot(pressure)
```
(ref:second-fig-caption) Here's a second complicated figure caption, also referencing Table \#ref(tab:mtcars).
```{r cars, fig.cap = '(ref:second-fig-caption)'}
plot(cars)
```
(ref:caption-table1) A caption for the first table. Check out this cross reference to Fig. \#ref(fig:pressure).
```{r mtcars}
mtcars |>
head() |>
kable(caption = '(ref:caption-table1)')
```
I can embed plots using just RMarkdown's {.tabset}
#### Heading {.tabset}
##### Subheading 1
```{r, echo=F}
df[[1]]
```
This produces individual tabs with the specified graphs (df is a list of graphs, calling df[[i]] produces a graph) in the preview pane (renders all the graphs inline in RStudio).
And I can generate just the tabs using a for loop.
```{r, results='asis', echo = FALSE}
for (i in 1:length(gg0)) {
cat("##### ",q$Subheading[i],"\n")
}
```
And this produces the desired output - the tabs with the names in the Subheading column.
However, I am stuck in trying to generate the graphs themselves using the for loop similar to how I did when I coded it manually.
Extending the above, I tried to generate the markdown that produced the initial output but the plot fails to generate (both in the inline markdown and preview).
```{r, results='asis', echo = FALSE}
for (i in 1:length(gg0)) {
cat("##### ",q$Subheading[i],"\n")
cat('```{r, echo=F} \n')
cat("gg0[[",i,"]]\n")
cat('``` \n')
}
```
Maybe I am missing a finer point regarding markdown? I have tried various patterns using cat (and even without)
I would prefer a RMarkdown solution but other solutions are just as welcome.
I played around a little and found a solution. You have to use print within the asis code chunk...
```{r}
library(ggplot2)
gg0 <- list()
gg0[[1]] <- ggplot(mtcars, aes(mpg, hp)) + geom_point()
gg0[[2]] <- ggplot(mtcars, aes(mpg, disp)) + geom_point()
gg0[[3]] <- ggplot(mtcars, aes(mpg, drat)) + geom_point()
headings <- c('hp','disp','drat')
```
#### Heading {.tabset}
```{r, results='asis', echo = FALSE}
for (i in 1:length(gg0)) {
cat("##### ",headings[i],"\n")
print(gg0[[i]])
cat('\n\n')
}
```
As an explanation, the cat command together with results='asis' produces the markdown code for a lower level headline and prints the ggplot graph afterwards. Since we used `{.tabset} in the parent headline, it creates the plots in separate tabs.
Adding both plot.new(), dev.off() inside the for loop solves the problem of adding all the figures in the last tab. See the complete solution here.
I label my figures like this.
---
title: "xxx"
output:
pdf_document:
fig_caption: true
---
And then in each chunk
```{r, fig.cap="some caption"}
qplot(1:5)
```
This works quite nicely. However in chunks where I plot multiple figures within a loop I can't specify a caption. This produces no caption at all:
```{r, fig.cap="another caption"}
qplot(1:5)
qplot(6:10)
```
How can I specify a figure that counts from the same number as the first chunk for each plot?
You can use a fig.cap argument of length 2 (or the size of your loop):
```{r, fig.cap=c("another caption", "and yet an other")}
qplot(1:5)
qplot(6:10)
```
Found an easy way to dynamically produce plots and add them to the pdf with individual captions, using knitr::fig_chunk as described here. This is also a workaround for OPs comment that message=false (or echo=False or results='asis' for that matter) supresses the fig.cap argument.
```{r my-plots, dev='png', fig.show='hide', echo=FALSE}
# generate plots first
qplot(1:5)
qplot(6:10)
```
```{r, echo=FALSE, results='asis'}
# then put them in the document with the captions
cat(paste0("![some caption](", fig_chunk(label = "my-plots", ext = "png", number = 1), ")\n\n"))
cat(paste0("![another caption](", fig_chunk(label = "my-plots", ext = "png", number = 2), ")\n\n"))
```
Hopefully this helps someone who stumbles upon this question in the future.
I am trying to use the R-Markdown feature in R Studio where I was trying to print plots which are generated inside a function. This is a basic run down example of what I am trying to do.
**Test printing plots generated in a function**
================================================
``` {r fig.width=8, fig.height=4, warning=FALSE, eval=TRUE, message=FALSE, tidy=TRUE, dev='png', echo=FALSE, fig.show='hold', fig.align='center'}
dat <- data.frame(x=c(1:10),y=c(11:20),z=c(21:30),name=rep(c("a","b"),each=5))
library(ggplot2)
ex <- function(data){
plot(data[,1],data[,2])
plot(data[,1],data[,3])
}
for (i in 1:10){
t1 <- rbind(i,ex(dat))
}
t1
```
Those testing this code, please make sure to save it as ".Rmd" file and then run the knithtml() in RStudio toolbar. This code above works absolutely fine with the kind of html output I desire. However when I replace the base plotting function by ggplot based code, I cannot get the knithtml() to produce the ggplot output of the 10 plots that I got like before. The base plot code above is now replaced by the following code
p1 <- ggplot(data=data, aes(x=data[,1],y=data[,2]))
p1 <- p1+geom_point()
p1
Am I missing something very simple here.
VJ
There are two problems in your code:
ggplot doesn't recognize the data x and y data, bacause it works inside the data environment. You should give it the column names directly.
The code in yur loop doesn't make sense. You can't mix a plot with an index... (the reason it works with the base plot is through a side-effect) I've replaced it with the simple plot command.
The following will work:
**Test printing plots generated in a function**
================================================
``` {r fig.width=8, fig.height=4, warning=FALSE, eval=TRUE, message=FALSE, tidy=TRUE, dev='png', echo=FALSE, fig.show='hold', fig.align='center'}
dat <- data.frame(x=c(1:10),y=c(11:20),z=c(21:30),name=rep(c("a","b"),each=5))
library(ggplot2)
ex <- function(data){
p1 <- ggplot(data=data, aes(x=x,y=y))
p1 <- p1+geom_point()
return(p1)
}
for (i in 1:2){
plot(ex(dat))
}
```