knitr and car::recode

[mk9y]

The following code:

<<getHowell>>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@

produces:

\begin{knitrout}
\definecolor{shadecolor}{rgb}{0.749, 0.937, 1}\color{fgcolor}\begin{kframe}
\begin{alltt}
> add <- \hlfunctioncall{read.table}(\hlstring{"howell.txt"}, header = TRUE)
> add$CaseNum <- \hlfunctioncall{factor}(add$CaseNum)
> \hlfunctioncall{library}(car)
> add$Gender <- \hlfunctioncall{recode}(add$Gender, \hlstring{" 1 = \hlstring{'Male'}; else = \hlstring{'Female'} "}, as.factor.result = TRUE)
> add$Repeat <- \hlfunctioncall{recode}(add$Repeat, \hlstring{" 1 = \hlstring{'RepeatedGrade'}; else = \hlstring{'DidNotRepeatGrade'} "}, 
       as.factor.result = TRUE)
> add$EngL <- \hlfunctioncall{recode}(add$EngL, \hlstring{" 1 = \hlstring{'CollegePrep'}; 2 = \hlstring{'General'}; else = \hlstring{'Remedial'} "}, 
       as.factor.result = TRUE)
> add$EngG <- \hlfunctioncall{recode}(add$EngG, \hlstring{" 4 = \hlstring{'A'}; 3 = \hlstring{'B'}; 2 = \hlstring{'C'}; 1 = \hlstring{'D'}; else = \hlstring{'F'} "}, 
       as.factor.result = TRUE)
> add$SocProb <- \hlfunctioncall{recode}(add$SocProb, \hlstring{" 0 = \hlstring{'No'}; else = \hlstring{'Yes'} "}, as.factor.result = TRUE)
> add$Dropout.f <- \hlfunctioncall{recode}(add$Dropout, \hlstring{" 1 = \hlstring{'Dropped'}; else = \hlstring{'DidNotDrop'} "}, 
       as.factor.result = TRUE)
\end{alltt}
\end{kframe}
\end{knitrout}

with throws the followeing error in the console:

! Undefined control sequence.
<argument> '
            Male'
l.1431 ...g{'Male'}; else = \hlstring{'Female'} "}
                                                  , as.factor.result = TRUE)

[yihui]

I see the line number is 1431, so it must be a very long document. I cannot reproduce the error with the following minimal document:

\documentclass{article}
\begin{document}
<<getHowell>>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@
\end{document}

Perhaps that is a LaTeX problem from elsewhere...

[mk9y]

It is indeed a huge document, but the reason I thought I had found the culprit is that the problem goes away when I comment out this block of knitr-generated LaTeX.
In any event, thanks so much for the helpful hint. I was also unable to reproduce using your example. I'll check further.

On Oct 29, 2012, at 3:54 PM, Yihui Xie notifications@github.com wrote:

I see the line number is 1431, so it must be a very long document. I cannot reproduce the error with the following minimal document:

\documentclass{article}
\begin{document}
<>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@
\end{document}
Perhaps that is a LaTeX problem from elsewhere...

—
Reply to this email directly or view it on GitHub.

[mk9y]

The conflict is with \usepackage{lmodern}

The following fails:

\documentclass{article}
\usepackage{lmodern}
\begin{document}
<>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@
\end{document}

A request: knitr() add a lot of LaTeX code in the preamble. How about writing a file with a different suffix that is included at the beginning of the .tex file.
And a question: as you saw, I'm writing a large document (with chapters). I would like to compile the Rnw one file at the time. Would a \knitrincludeonly{} command be possible? This would also insure that there are no inter-chapter dependencies.

[yihui]

Thanks for digging out the culprit and the suggestions.

You can probably try opts_knit$set(self.contained = FALSE) so the large preamble goes to a file knitr.sty.

For the second question, I think you can use the set_parent() function in your chapters; see http://yihui.name/knitr/demo/child/

[yihui]

\usepackage{lmodern} works fine for me with TeXLive 2012 under Ubuntu 12.10.

[mk9y]

Strange. I keep my LaTeX updated weekly (OS X 10.8.2).

MK

On Oct 29, 2012, at 9:51 PM, Yihui Xie notifications@github.com wrote:

\usepackage{lmodern} works fine for me with TeXLive 2012 under Ubuntu 12.10.

—
Reply to this email directly or view it on GitHub.

[mk9y]

I have another question. Can I put some of the knitr code into a boxedminipage (e.g., for exercises). The LaTeX compilation fails:

\documentclass{article}
\usepackage{boxedminipage}
\begin{document}
<>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@
\boxedminipage{
<>=
add <- read.table('http://www.uvm.edu/~dhowell/methods7/DataFiles/Add.dat', header = TRUE)
add$CaseNum <- factor(add$CaseNum)
library(car)
add$Gender <- Recode(add$Gender, " 1 = 'Male'; else = 'Female' ", as.factor.result = TRUE)
add$Repeat <- Recode(add$Repeat, " 1 = 'RepeatedGrade'; else = 'DidNotRepeatGrade' ", as.factor.result = TRUE)
add$EngL <- Recode(add$EngL, " 1 = 'CollegePrep'; 2 = 'General'; else = 'Remedial' ", as.factor.result = TRUE)
add$EngG <- Recode(add$EngG, " 4 = 'A'; 3 = 'B'; 2 = 'C'; 1 = 'D'; else = 'F' ", as.factor.result = TRUE)
add$SocProb <- Recode(add$SocProb, " 0 = 'No'; else = 'Yes' ", as.factor.result = TRUE)
add$Dropout.f <- Recode(add$Dropout, " 1 = 'Dropped'; else = 'DidNotDrop' ", as.factor.result = TRUE)
@
}
\end{document}

On Oct 29, 2012, at 9:51 PM, Yihui Xie notifications@github.com wrote:

\usepackage{lmodern} works fine for me with TeXLive 2012 under Ubuntu 12.10.

—
Reply to this email directly or view it on GitHub.

[yihui]

I guess it is a bad idea in general to put complicated content in \command{} due to the potential problem of } in the content, which can confuse LaTeX. The safer way is to use environments. I just tested boxedminipage environment and it works fine:

\begin{boxedminipage}{\linewidth}
<<>>=
# code here
@
\end{boxedminipage}

[yihui]

Do you have further questions on this issue? If not, I'll close it. Thanks!

[mk9y]

Please close.

MK

On Nov 13, 2012, at 1:02 AM, Yihui Xie notifications@github.com wrote:

Do you have further questions on this issue? If not, I'll close it. Thanks!

—
Reply to this email directly or view it on GitHub.

[github-actions]

This old thread has been automatically locked. If you think you have found something related to this, please open a new issue by following the issue guide (https://yihui.org/issue/), and link to this old issue if necessary.