[knitr] Options: Chunk options and package options | /knitr/options/

[yihui]

Options: Chunk options and package options

https://yihui.org/knitr/options/

[giscus-bot]

Guest *MartinInFfm* @ 2012-03-23 11:13:32 originally posted:

When I use xtable in a chunk, I  get a grey shaded empty line.
What can I do to suppress it?
I have tried already the options : warning=FALSE,message=FALSE,error=FALSE

Example:
<<print,echo=false,results='asis',warning=false,message=false,error=false>>=
print(xtable(data.frame(x=1)))
@

[yihui]

Can you try the latest version (0.4)? I do not see the line with 0.4.

Originally posted on 2012-03-24 05:31:26

[giscus-bot]

Guest *MartinInFrankfurt* @ 2012-03-24 08:04:37 originally posted:

It is gone now.
Thank you, Yihui!

[giscus-bot]

Guest *Bastien Ferland-raymond* @ 2012-04-03 17:26:32 originally posted:

Hello Yihui,

First, thanks a lot for your work, this package is great!

I have a problem adding a figure using the layout() function and was wondering if you could help me.

An exemple code:

<<virees, results="hide" ,echo="F,fig=T,dev="png",dpi=150,fig.height=6,fig.width=10,fig.keep="high"," fig.align="center" ,fig.height="6,fig.width=10">>=
for(i in c("g4564", "g4167")){
windows(10,6)
layout(matrix(c(1,2),1,2), widths=lcm(c(25.4-15.24,15.24)), heights=lcm(c(15.24,15.24)))
par(mar=c(0,0,0,0))
plot(1,1); text(1.1,1,i)
plot(2,1, col="red"); text(2.1,1,i)
dev.off()
}

If I use that, my figures include only the left plot in the layout.  Any idea why?

I actually want to print in a report with multiple figures (~250) where each one is the result of a special layout format.

Thanks a lot

Bastien

[giscus-bot]

Guest *Bastien Ferland-raymond* @ 2012-04-04 12:10:46 originally posted:

 Thanks for your reply,

The reason I've put windows(10,6) is because that is the only way I've found it would make the require graph.  if I do not put that there, I get and error in R "Error in plot.new() : figure region too large" which then print in my pdf file.  It's like if putting fig.height=6 and fig.width=10 doesn't make a device of those dimension but instead uses the system options("device").

Using windows(10,6) without the loop works (so making only one figure).
Using windows(10,6) with the loop (or I also tried with sapply()) doesn't print the full graphics
Not using windows(10,6) doesn't print anything because of device size issues.

Any idea?

Bastien

[yihui]

If you are using the latest version of knitr and running it under non-interactive mode (e.g. call Rscript), this should not be a problem; knitr will use fig.width and fig.height to open options('device').

If you are in an interactive R session (open R and type code), this might be problematic. It is easy to fix either on my side, though. You can also fix this by yourself by putting something like this in your first chunk:

options(device = function(...) {
  windows(..., width = opts_current$get('fig.width'), 
          height = opts_current$get('fig.height'))
})

However, I still recommend you to call R non-interactively.

Originally posted on 2012-04-04 17:29:57

[giscus-bot]

Guest *Bastien Ferland-raymond* @ 2012-04-04 19:01:07 originally posted:

 Thanks again for your time, it's very appreciated.

There is either something that I don't understand, or a bug somewhere.  I'll try to give you as much information as possible, it may help you figure out what the problem is.

I'm not sure I understand very well what does using R in non-interactive mode mean, but I think it's what I'm doing.  I use Tinn-R to do my coding from which I send my code.  I write my rnw file and I call it using:

setwd("Z:rapport latex")
knit("test1 rapport sonar.rnw")
texi2dvi("test1 rapport sonar.tex",pdf=T,texinputs = "R_HOME/share/texmf", quiet=T, clean=T)

I've downloaded your package on monday at the same time as R 2.15.0, so I guess I have the last version.  I'm using win XP 32bit.

I've tried to had the code you gave me in the previous post in my first chunk.  However, I got an error:
"Quitting from lines 79-81: Error in getOption("device") : node stack overflow"

Line 79-81 represent my second chuck which has the options:
<<nppt, echo="F," dev="png" ,fig="T,fig.height=4,fig.width=7," fig.cap="Nombre de placettes potentielles par tuile" ,fig.lp="placpartuile" ,out.width=".4linewidth" , ="" fig.align="center" ,fig.pos="H">>=

However, if I run the code again without changing anything, it give me the same error but for the first chunk. It's like it keeps the same options and doesn't like it.

Also, when you say: "knitr will use fig.width and fig.height to open options('device')", that is exactly the way I expect it to behave, but I douth it's doing it, on my computer at least.  If the opened device was the good size, I wouldn't get a "figure region too large" error.  And also, it may be unimportant, but when I run knit(), I see my figures flashing on my screen and they are all the default 7in x 7 in, and not size I've previously define (however, the final file saved are the proper size).

what do you think?  Do you need more info?

thanks

Bastien

[yihui]

Sorry, I had a typo in the code above; dev.new() should be windows(). For interactive vs non-interactive sessions, see ?interactive in R help.

Anyway, I have fixed this problem in the development version, so you do not really need to do anything but install the latest version from Github: https://github.com/yihui/knitr (or wait for my next CRAN release). Thanks!

Originally posted on 2012-04-04 22:40:51

[giscus-bot]

Guest *Bastien Ferland-raymond* @ 2012-04-05 14:56:03 originally posted:

 Great!  I'll wait for the cran release and let you know if I run into more trouble.

Thanks a lot,
Bastien

[giscus-bot]

Guest *Morales* @ 2012-04-08 08:21:02 originally posted:

Hi Yihui, congratulations for your work with knitr. I am using it a lot, but I am having problems when I try to supress some lines in the output. It's fine when I use i.e. "echo=2", but I am not able to get "echo= 1:3" . Try this example please:
 begin{document}
<<p1,echo=f,comment=na>>=
rendimiento = c(30,34,25,21,24,11,25,12,9)
metodo = factor(c(1,1,1,2,2,2,3,3,3),
labels= c("A","B","C"))
mod=aov(rendimiento~metodo)
@
<<p2,echo=1:3,comment=na,tidy=f, results="asis">>=
library(multcomp)
tuk1 = glht(mod, linfct= mcp(metodo="Tukey"))
res=summary(tuk1)
x1a=res$test$coefficients
x2a=res$test$tstat
x3a=res$test$sigma
x4a=res$test$pvalues
a1=round(data.frame(x1a,x3a,x2a,x4a),3)
colnames(a1)=c("Estimación","Std","T","P")
library(xtable)
print(xtable(a1,caption="Comparaciones: Método Tukey"),caption="top",floating=T)
@
end{document}

[giscus-bot]

Guest *Morales* @ 2012-04-08 08:36:53 originally posted:

Hi Yihui,forget my previous post. I forgot the quotation marks in results= "asis" . Thanks for your fine work

[yihui]

Good to know. You should see a warning message about this situation when you call knit().

Originally posted on 2012-04-08 18:53:44

[giscus-bot]

Guest *jamiefolson* @ 2012-05-15 14:47:45 originally posted:

What is the best practice for chunks where you only want the figures?  I can turn off all of the output, but that still leaves a colored, framed box around all the chunks, even if it's just an includegraphics.  I also end up with lots of stuff added to the main document header that's never used.  Is there any way you could just put that stuff in a latex package and include it like framed is?

I can turn off the highlighting, but I still get the framed colorbox.

[giscus-bot]

Guest *Eric Nord* @ 2012-05-15 15:23:16 originally posted:

 I'm trying to figure that out also.

[yihui]

Normally you can use echo=FALSE, results='hide' to keep the figures only. If you still find the gray box, you probably need to update the package.

For the large preamble, currently the suboptimal solution is opts_knit$set(self.contained = FALSE).

Originally posted on 2012-05-15 19:32:26

[giscus-bot]

Guest *Eric Nord* @ 2012-05-15 18:01:54 originally posted:

Does anyone have nice example of using chunk options in markdown with knitr? I'm lost.
 

[yihui]

Most examples are under https://github.com/yihui/knitr/tree/master/inst/examples ; in your case, you can look at knitr-minimal.Rmd or knitr-upload.Rmd; actually these chunk options are exactly the same as the ones you write in Sweave. Instead of <>=, you just write ``` {r options}.

Originally posted on 2012-05-15 19:13:41

[giscus-bot]

Guest *Florian Oswald* @ 2012-05-25 01:51:30 originally posted:

hi yihui!
I recently updated to the latest knitr version (along with R 2.15). I had a very nice (and fast) running setup with lyx before the upgrade. now knitr complains that i am using old Sweave options and I should visit this page. I'm having a hard time figuring out which options are outdated and which ones are not. Is there a kind of list for this somewhere? I guess it's slower because of this. Thanks for your help!
Florian

[giscus-bot]

Guest *Florian Oswald* @ 2012-05-25 01:57:42 originally posted:

wait a moment. just discovered the transitions from sweave page. need to write 
results='hide'
now instead of 
results=hide
 that was a useless question, I guess. 
see you around for more useful questions, I hope!
Florian

[yihui]

Glad you figured it out by yourself. All the valid option values are listed in this page, and I have emphasized in the beginning of this page that options must be valid R code.

Originally posted on 2012-05-25 02:00:37

[giscus-bot]

Guest *Eugene* @ 2012-05-30 19:52:55 originally posted:

Hi Yihui,

I am running Lyx 2.0.3 on Windows XP with loaded module Knitr and have following error popping up while trying to convert R code to TeX:

"An error occurred while running>
Rscript --verbose --no-restore "C:/Program Files/LyX20/bin/../Resources/scripts/lyxknitr.R" "C:/Documents and Settings/me/Local Settings/Temp/lyx_tmpdir.Hp3588/lyx_tmpbuf2/""Paper2.Rnw""C:/Documents and Settings/me/Local Settings/Temp/lyx_tmpdir.Hp3588/lyx_tmpbuf2/""Paper2.tex" ISO-8859-15"M:/CODE/Paper2/" 

"Do you have a clue what it might be?My document have an empty preamble. I only use the following options in the beginning of the document"
<<set-options, include="FALSE," echo="FALSE," cache="FALSE">>=options(replace.assign=TRUE,width=90)knit_hooks$set(par=function(before, options, envir){if (before) par(mar=c(4,4,.1,.1),cex.lab=.95,cex.axis=.9,mgp=c(2,.7,0),tcl=-.3)})read_chunk('Paper2.R')opts_chunk$set(fig.path='figure/', cache.path='cache/', dev='tikz', external=TRUE, fig.width=7, fig.height=4, fig.align='center', fig.show='hold',cache=FALSE, par=TRUE) at "
If i run the document with R chunks without bold options, it seems to work, although output needs aligning and sizing.

In R i have loaded both knitr and tikzDevice packages. I wrote "at" instead of a symbol only here due to mail autocompletion...

Best regards,
Eugene

[giscus-bot]

Guest *Eugene* @ 2012-06-01 07:46:52 originally posted:

I realized that bold text disappeared and lines are merged. I meant opts_chunk$set()

[giscus-bot]

Guest *Eugene* @ 2012-06-01 08:52:19 originally posted:

The LyX converts file with no problems if i change output device from tikz to png or pdf. 

"opts_chunk$set(fig.path='figure/', cache.path='cache/', dev='tikz', external=TRUE, fig.width=7, fig.height=4, fig.align='center', fig.show='hold',cache=FALSE, par=TRUE) "

Why LyX 2.0.3 cannot convert R file if i select tikz device?

[giscus-bot]

Guest *Eugene* @ 2012-06-01 10:32:42 originally posted:

I also experience a problem with automatic update of images. They are properly updated only if content of a folder FIGURE is cleared. What can it be?

[yihui]

take a look at the corresponding TeX log file under figure/ to see what is wrong there; tikz requires LaTeX expertise, so if you are not familiar with the tikzDevice package, you probably should spend a couple of hours learning it.

Originally posted on 2012-06-01 20:08:25

[yihui]

LyX will not really update your document unless there has been a change in the TeX output; if you only update a plot, the tex code is probably not changed although the plot itself has been changed. In that case, you need to make a small change in any place of your LyX document.

Originally posted on 2012-06-01 20:12:26

[giscus-bot]

Guest *Rafik Margaryan* @ 2012-07-04 07:45:13 originally posted:

Hi Yihui,
I've got confused with references and it seems does not work for me. If I put chunck label for plot fig1, how shoud referre to that figure by default : fig:fig1 or just fig1?
And how should I refer to tables( I use Hmisc summary.formula for creating tables)?
Many thanks

[yihui]

Because fig.lp is 'fig:' by default, you need to ref{fig:fig1} instead of ref{fig1}. If you want the latter, you can set opts_chunk$set(fig.lp = '') so that no prefix will be added by knitr.

Originally posted on 2012-07-04 20:55:14

[giscus-bot]

Guest *Chris Okugami* @ 2012-07-09 20:48:45 originally posted:

Hi, Yihui, 

I got a problem using kintr and markdown on ubuntu. x64

here is my markdown code. 
mytest.Rmd 

this code generates animation buttons but it doesn't play on web browser. 

{r echo=FALSE,fig.path='figure/fig2',fig.keep='all',[fig.show](http://fig.show)='animate'}  plot(rnorm(100))plot(rnorm(100))
yes, i have mp4 support on ubuntu, with internet connection.

Any idea? 

[yihui]

I have heard other users report the same problem under Ubuntu, and I do not know why it does not work for you. I'm using Ubuntu as well.

There was indeed a small bug before v0.6.3, just in case you were using an older version.

Originally posted on 2012-07-10 01:12:27

[giscus-bot]

Guest *Chris Okugami* @ 2012-07-10 01:24:36 originally posted:

FYI: 
I am using  v0.6.3 knitr version. 

[yihui]

You probably need par(bg='white') if you are using base R graphics; I just found that FFmpeg did not work with PNG's with transparent background.

Originally posted on 2012-07-10 16:07:13

[giscus-bot]

Guest *Chris Okugami* @ 2012-07-11 23:59:28 originally posted:

Thanks, Yihui,
par(bg='white') did produced a proper video mp4 file, however, when I try to view the generated context, failed to load on the browser. I've check my web browser ensure of html5 video on my chrome and firefox, seems to be ok - facebook works so. my suspision is the type of mp4 video codec, which is not common to most of all computing environment. What do you think?   

[yihui]

MP4 should be very common nowadays. I have no idea about what is wrong. Anyway, you will soon be able to write animations into HTML pages like those in http://taiyun.github.com/

Originally posted on 2012-07-12 03:55:56

[giscus-bot]

Guest *Chris Okugami* @ 2012-07-09 21:00:55 originally posted:

animation option. 

Currently, support animation.fun ="hook_ffmpeg_html". 

Is it possible to use animated gif as  output? instead of mp4? 

It will be very nice to make animated PDF from markdown so that we can space in our report!! 

My motivation behind this is that 
I work with large dataset, many sampling points, very tiring of scrolling down!!

Thank for your help
Chris  
 

[yihui]

It is almost impossible to convert animations in HTML to PDF, because they were drastically different in this aspect. That being said, you can create either LaTeX/PDF animations or HTML animations with the same code and syntax (fig.show='animate'), so you'd better do that from the source rather than generating markdown and convert the animations to PDF later.

It is certainly possible to do GIF. Just file a feature request to https://github.com/yihui/knitr/issues

In the development version of knitr, you can already make animations with JavaScript (hook_scianimator) or Flash (hook_r2swf), but no instructions have been given. We are working on it. Hopefully you will see some interesting announcements next week.

Originally posted on 2012-07-10 01:19:47

[giscus-bot]

Guest *Donlumor* @ 2012-07-10 08:32:49 originally posted:

senic ¡- read.table(”pdata.txt”, header = FALSE) colnames(senic) ¡- c(”y”, ”x1”, ”x2”, ”x3”, ”x4”,
”x5”, ”x6”, ”x7”, ”x8”, ”x9”, ”x10”) attach(senic) pairs(senic, pch = 21, bg = c(”red”, ”yellow”, ”blue”,)

I used this code for the above result
<<echo=true,fig=true,comment="", eval="TRUE">>=senic<-read.table("pdata.txt",header=FALSE)colnames(senic) <- c("y","x1","x2","x3","x4","x5","x6","x7","x8","x9","x10")attach(senic)pairs(senic, pch=21, bg=c("red","yellow","blue","black"))

My output does not execute line by line
I used <- to store  read .table but i get  ¡- in output

I use Rstudio for my reports

[yihui]

I do not understand your question. I need a reproducible example along with your sessionInfo() after library(knitr).

BTW, fig=TRUE has been deprecated long time ago; you do not need fig=TRUE in knitr.

Originally posted on 2012-07-10 16:05:54

[giscus-bot]

Guest *Donlumor* @ 2012-07-10 16:52:57 originally posted:

documentclass{article}author{Richard LumorSTAT 482: REGRESSION ANALYSIS}title{MODELLING THE AVERAGE LENGTH OF STAY IN A HOSPITAL USING SENIC DATA}usepackage{amsmath}usepackage{setspace}usepackage{multirow}usepackage{graphicx}usepackage{times}usepackage{float}usepackage{verbatim}usepackage{geometry}begin{document}SweaveOpts{concordance=TRUE}maketitlesection{Introduction}We are going to do an exploratory analysis of the SENIC hospital data. We will use both qualitative and quantitative techniques.<<echo=true,comment="">>=senic<-read.table("pdata.txt",header=FALSE)colnames(senic) <- c("y","x1","x2","x3","x4","x5","x6","x7","x8","x9","x10")attach(senic)pairs(senic, pch=21, bg=c("red","yellow","blue","black")) @the plot shows correlation between $x_{8}$ and $x_{10}$&lt;<echo=true,comment="">>=cor(senic)boxplot(senic)@A look at the boxplot shows that <<echo=true>>=lm.senic <- lm(y～x1+x2+x3+x4+x5+factor(x6)+factor(x7)+x8+x9+x10)summary(lm.senic)@we will find the confidence and prediction intervals of the choosen model. The 95% confidence interval is given by <<echo=true,comment="">>=confint(lm.senic,level=1-0.05/7)@section{Model Selection}Some of the methods which will be used for our model selection process are Bayesian Information Criteria (BIC), Akaike Information Criterion (AIC), Mallows $C_{p}$ and $R_{adj}^2$.<<echo=true,comment="">>=library("lmtest")library("car")dwtest(lm.senic)ks.test(lm.senic$res, "pnorm", mean=0, sd=sd(lm.senic$res))bptest(lm.senic)boxplot(lm.senic$res)@the quantile plot is given by<<echo=true,comment="">>=qqnorm(lm.senic$res)qqline(lm.senic$res)@we will use the the steps function to determine which model will have the best $R_{adj}^2$&lt;<echo=true,comment="">>=library("leaps")temp <- regsubsets(y～x1+x2+x3+x4+x5+factor(x6)+factor(x7)+x8+x9+x10,data=senic)steps <- summary(temp)steps@<<echo=true,comment="">>=cbind(AdjR2=steps$adjr2, steps$which)cbind(Cp=steps$cp, steps$which)@section{MODEL SPECIFICS}The candidate model is $Y_{}=beta_{o}+beta_{1}X_{1}+beta_{2}X_{2}+beta_{7i}(factorX_{7i})+X_{8}+X_{9} + epsilon_{i}$ where $i=2,3,4$ for $X_{7i}$begin{center}begin{tabular}{llll}Name/Variable & $X_{7,2}$ & $X_{7,3}$ & $X_{7,4}$  NE & 0 & 0 & 0  NC & 1 & 0 & 0  S & 0 & 1 & 0 W & 0 & 0 & 1 end{tabular}end{center} 

library(knitr)
sessionInfo()

library(knitr)> sessionInfo()R version 2.15.1 (2012-06-22)Platform: x86_64-pc-mingw32/x64 (64-bit)locale:[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   [3] LC_MONETARY=English_United States.1252 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] knitr_0.6.3loaded via a namespace (and not attached):[1] digest_0.5.2   evaluate_0.4.2 formatR_0.5    plyr_1.7.1     stringr_0.6   [6] tools_2.15.1  

[yihui]

Oh, this is not an appropriate place to paste the code. I can barely read it. Could you please file an issue to https://github.com/yihui/knitr/issues ? Thanks! And I doubt if the above example is reproducible for me, because I do not have pdata.txt

Originally posted on 2012-07-10 20:20:23

[giscus-bot]

Guest *Sebastian Campbell* @ 2012-07-13 04:49:10 originally posted:

testfun.Rmd:

summary(cars)  

I'm not sure if this is a bug (and it doesn't cause knitting a document to stop), but this causes and error:

options(encoding=''); require(knitr); knit('testdash.Rmd');
Loading required package: knitr

processing file: testdash.Rmd
Error in parse(text = str_c("alist(", params, ")")) :
1:20: unexpected ','
1: alist( car-function,
^
Warning in parse_params(params) :
(*) NOTE: I saw options " car-function, eval=FALSE"
are you using the old Sweave syntax? go http://yihui.name/knitr/options

[yihui]

I did not mention this in the documentation. Strictly speaking, chunk labels must be quoted as well because they are character strings. In most cases they do not need to be quoted because they are valid R expressions if not being quoted, e.g. a-b, or c_def. However, in your case car-function is not a valid R expression, because "function" is a reserved keyword in R. I will add an explanation to this page later. Thanks for the report!

Originally posted on 2012-07-13 17:26:35

[giscus-bot]

Guest *Gurix* @ 2012-07-17 16:49:01 originally posted:

Good work bro. That's indeed a significant step forward in R history.

[giscus-bot]

Guest *wegelinjacob* @ 2012-07-18 19:45:13 originally posted:

At http://jacobwegelin.net/tmp/XieYihui/, in the 
20120718Wed151045-foo.Rhtml example, I specified "echo=FALSE" to try to show only the results. But when the resulting html file (also on the website) is displayed in a browser, a bunch of pound or number signs (#) are displayed in place of the R source code. Is there a clean automatic way to get rid of all traces of the R code in the displayed html, so that one only sees (in this case) the displayed table? 

The pertinent documentation (from http://yihui.name/knitr/options) is: 

echo: (TRUE; logical or numeric) whether to include R source code in the output file

In the same place a 2nd html file can be found which contains the lines from the prevous html that are necessary to display the table, without the preamble that starts the first html file.  What is the function of those pound signs?  

[yihui]

use results='asis'; because you are writing raw HTML from R

Originally posted on 2012-07-18 20:51:52

[giscus-bot]

Guest *George Flynn* @ 2018-08-18 08:01:12 originally posted:

Yihui - hello - I am new to knitr and am trying to create a PDF document for an assignment, I am getting the following error:
processing file: Cars.Rmd
|...... | 9%
ordinary text without R code

|............ | 18%
label: setup (with options)
List of 1
$ include: logi FALSE

|.................. | 27%
ordinary text without R code

|........................ | 36%

Quitting from lines 25-32 (Cars.Rmd)
Error in eval(x, envir = envir) : object 'False' not found
Calls: ... process_group.block -> call_block -> eval_lang -> eval -> eval
Execution halted

Lines 25 -32 are:

library(knitr)  
library(ggcorrplot)  
library(ggplot2)  
data(mtcars)

corr <- round(cor(mtcars), 2)  
ggcorrplot(corr, hc.order = TRUE, type = "lower",lab = TRUE)  

I am assuming it has something to do with the chunk options?
Thanks George

[giscus-bot]

Guest *Matthias* @ 2019-03-23 15:43:16 originally posted:

Yihui, great work - I love to use knit, actually with .Rnw to write a textbook.

One question: Is there a way to remove the empty line between the function call and the respective output of the function?

Thanks a lot for your support!

[giscus-bot]

Guest *Pedro Torres* @ 2020-02-06 01:01:10 originally posted:

Whenever i print a large vector or a frame, i always get part of my R output out of the margins... ¿Is it possible to adjust the output margins?

[giscus-bot]

Guest *Jon Wakefield* @ 2020-07-06 22:14:51 originally posted:

I use beamer slides extensively, as created by Rmarkdown, and always have problems with R output being too wide.
Can you help me with the following simplified file:

---  
title: "Test File"  
author: |  
| Jon  
| University of Washington  
date: "`r Sys.Date()`"  
output:  
beamer_presentation:  
theme: "AnnArbor"  
colortheme: "dolphin"  
keep_tex: yes  
slidy_presentation: default  
---

```{r setup, echo=FALSE, message=FALSE, warning=FALSE}  
library(knitr)  
opts_chunk$set(collapse=TRUE, fig.align='center', tidy=TRUE, tidy.opts=list(blank=TRUE, width.cutoff=20,strip.white=TRUE), warning=FALSE,message=FALSE,cache=TRUE)  
```  
## test

```{r, message=FALSE, collapse=TRUE, tidy=TRUE,tidy.opts=list(width.cutoff=30)}  
seq(1,100)  
```

[giscus-bot]

Guest *ERWIN LARES* @ 2021-02-20 19:30:43 originally posted:

Hello Smart People!

Here with a question about cross-referencing. I am successfully using @ref(tab:my-table) and @ref(fig:my-figure) to cross reference tables and figures that I created inside R chunks. Yay for me.

Then I started writing numbered linguistics examples with code like this:

begin{exe}  
exlabel{stm-01}  
hfil  
begin{tikzpicture}[baseline=0pt]  
Tree [.{textbf{Aspectual Classes}} [.textbf{Statives} textit{States} ][.textbf{Dynamic} [.textbf{Atelic} textit{Activities} ][.textbf{Telic} [.textbf{Instantaneous} textit{Achievements} ][.textbf{Durative} textit{Accomplishments} ]]]]  
end{tikzpicture}  
hfil  
medskip  
end{exe}  

When I was trying to cross reference my example I was stuck. I trying adding a name to the chunk and captioning the chunk, but no avail. Then by chance I wrote this ref{stm-01} and it work like a charm.

So my question is, what is the difference between @ref(tab:my-table) and ref{my-label}

Much appreciative!

[giscus-bot]

Guest *Ale Campoy* @ 2021-03-07 00:51:18 originally posted:

Why is this happening? I cannot find the proper argument. My Rmd is:

---

title: |
| Title
output:
pdf_document: default
html_document: default
always_allow_html: 'true'

knitr::opts_chunk$set(fig.pos = 'H' )  
knitr::opts_chunk$set(echo = TRUE)  

Paquetes que vamos a usar en alguno de los ejercicios:

library(rio)  
library(rknn)  
library(kknn)  

0. Análisis exploratorio y transformación de los datos

Primero cargamos el archivo con los datos

setwd("P:/MDS/AEM/Entrega 2") # Mi Working directory, cambiar para que funcione  
dataset_raw <- import("khantrainadd.csv")  
dataset_raw$V1 <- NULL # esta columna con el indice no nos interesa  
head(dataset_raw)[,1:5]  
table(dataset_raw$class) # Mi variable objetivo  
dim(dataset_raw)  

and I get inverted order of the chunks, no matter what I do the following pdf knit https://uploads.disquscdn.com/images/e716a777b15a9bbe91aea5ebb629b1f485af61922031eeaffb6eea31453ecc8b.png

[giscus-bot]

Guest *Ming* @ 2021-04-11 12:33:57 originally posted:

Hi Yihui, in R notebook inline display, the real plot size in the chunk display is consistently about 2.5-3 times larger than the specified fig.width and fig.height in the chunk options. Any comments on what might be causing this issue? Plus, the font size in the chunk display is about 50% larger than the specified font size, when compared with ggsave output. I have tried updating my Rstudio, R, didn't change anything. I also tried to use external monitors, same behavior. I use R notebook as a workhorse and having the right size plot in the chunk display is really important to me. Any insight would be appreciated!

[yihui]

That sounds like an RStudio issue, which is beyond my expertise. Please consider asking the question on RStudio Community or file an issue to the Github repo rstudio/rstudio. Thanks!

Originally posted on 2021-04-12 13:39:30

[giscus-bot]

Guest *rt* @ 2021-11-08 06:12:46 originally posted:

~ code chunks are related to links on a page while r relates to data packages

[yihui]

Sorry, I didn't get it. What do you mean?

Originally posted on 2021-11-12 05:46:53

[giscus-bot]

Guest *Mike Gilchrist* @ 2021-11-10 05:29:16 originally posted:

I'm using knitr with emacs and ESS. My console output is getting truncated with with the text "[TRUNCATED]". I believe this is the result of knitr. Is there a setting I can change to remove this?

[yihui]

This is not caused by knitr; knitr doesn't have control over the R console.

Originally posted on 2021-11-10 14:25:09

[giscus-bot]

Guest *Mike* @ 2021-11-13 13:47:46 originally posted:

Okay. Thanks for clearing that up.

[giscus-bot]

Guest *peacecop kalmer:* @ 2022-05-11 18:29:04 originally posted:

How can i include LaTeX-package declarations for HTML output? For instance, if i choose PDF as output format, i can write "usepackage{...}" into a separate TEX-file that i reference in RMD-file by "template: ...". However, if i don't use a separate TEX-file and only want HTML-output then the built-in LaTeX-packages work fine and i can use equations using them but if i want to use a non-buil-in-package then what should i do? There are so many examples out there for primitive things but once it gets complicated i seldom find a solution.

[giscus-bot]

Guest *MrDJR* @ 2017-04-25 14:33:21 originally posted:

Sorry for the poor formatting. I can't seem to preview, edit, or delete. I'll try once more with escape characters:
Hi! I'm trying to append a custom class to a figure div by supplying class.output='myClass' to my chunk options as follows:

plot(1:10)  

I expect the page element to be <div class="figure myClass" ="">; but instead it remains <div class="figure" ="">. I'm rendering the page with bookdown::html_document2. What am I doing wrong?

[courtiol]

FYI: in

size: ('normalsize'; character) Font size of the chunk output from .Rnw documents. See this page for possible sizes.

The URL link is broken. Perhaps this is temporary(?), in the meantime, here is another page covering the topic: https://latex-tutorial.com/changing-font-size/

[yihui]

Thanks for the reminder! I can access the link right now. Perhaps it was a temporary problem.

[drag05]

Hello yihui

I am rendering a pdf document from within Shiny app and at one instance I need to include a png plot previously saved. The Rmd file is placed in the root directory of the shiny app along with global.R, server.R and ui.R. Function includegraphics() works fine if I use it inline but not so inside code chunk:

Inline, but with no caption (caption throws a "float" error in this case):

knitr::include_graphics('./src/coverage001.png', dpi = 72, auto_pdf = TRUE)

Inside code chunk preferably with caption:

```{r covPlot, fig.align='center',  fig.ncol = 1, fig.sep = '\\par', fig.width = 9, fig.show = 'asis', fig.keep = 'all', fig.cap = sprintf('Protein %s: Coverage', names(chain))}
# coverage plot
knitr::include_graphics('./src/coverage001.png', dpi = 72, auto_pdf = TRUE)
```

When I run the chunk in RStudio, the figure appears under the chunk but not so inside the rendered document. No latex error either in the log file. I am using graphicx package without any success and all R packages are at their latest version. I have also `tlmgr_reinstalled tinytex and updated, to no avail. Please advise, thank you!

[yihui]

the figure appears under the chunk but not so inside the rendered document

I'm not sure what that means. Do you know that figure environments in LaTeX float by default? That is, usually they don't appear exactly where you place them in the source document. Reference: https://bookdown.org/yihui/rmarkdown-cookbook/figure-placement.html

[drag05]

@yihui Thank you for the prompt reply!
Is not about where the figure appears in the pdf document. My question was referring to why the figure only appears in the pdf document when include_graphics is used inline and why it does not appear at all in the pdf document when include graphics is used inside the code chunk.

In other words, what should I add to the chunk options to have the figure appear when I use include_graphics inside the code chunk (please see chunk options in example 2).

This is what I mean by inline (example 1):

`r knitr::include_graphics('./src/coverage001.png', dpi = 72, auto_pdf = TRUE)`

and this is what I mean by code chunk (example 2):

```{r  fig.align='center',  fig.ncol = 1, fig.sep = '\\par', fig.width = 9, fig.show = 'asis', fig.keep = 'all', fig.cap = "A00B10"}
knitr::include_graphics('./src/coverage001.png', dpi = 72, auto_pdf = TRUE)
```

This is part of my document header, including latex packages (the indents are stripped by the browser - they are correct in the Rmd):

output: 
  pdf_document: 
    latex_engine: xelatex
    df_print: paged
    fig_caption: yes
    toc: yes
    extra_dependencies: float
    highlight: tango
lof: yes
lot: yes
fontawesome: yes
fig.align: center
graphics: yes
header-includes:
 \usepackage{fontspec} \usepackage{graphicx} \usepackage{pdflscape} \usepackage{rotating} \usepackage{titling} \setlength{\droptitle}{-1cm} \usepackage{float} \let\origfigure\figure \let\endorigfigure\endfigure \renewenvironment{figure}[1][2] {\expandafter\origfigure\expandafter[H]} {\endorigfigure} \usepackage{mdframed} \usepackage{subfig} \usepackage{fancyhdr} \setlength{\headheight}{42.45442pt} \addtolength{\topmargin}{-1.3pt} \usepackage{graphicx} \usepackage{booktabs,xcolor} \usepackage{lastpage} \usepackage{longtable} \pagestyle{fancy} \fancyhf{} \fancyfoot[R]{\thepage} \setlength\footskip{14.5pt} \fancypagestyle{plain}{\pagestyle{fancy}} \lhead{\today}  \rhead{\includegraphics[width=2.5cm]{./www/logo.png}}
mainfont: Palatino Linotype
geometry: margin = 0.75in
fontsize: 11pt

Regarding latex packages: where can I find some succinct information on what latex package(s) to call with \use{} for various tasks?

Thank you!

[yihui]

My question was referring to why the figure only appears in the pdf document when include_graphics is used inline and why it does not appear at all in the pdf document when include graphics is used inside the code chunk.

I see. That's odd. Sorry I have no idea... A minimal reproducible example will be helpful.

[guanghu52]

I had a render pdf problem when my images had multiple dots in the filename, knitr.include.graphics.ext not run?