Frequently Asked Questions

The FAQ's are compiled from the issues and messages I received from blog posts and emails, etc.

Please first update all your R packages (use update.packages()) and probably R itself (what is the current R version?), then see if it works; if not, file an issue to me (email is fine, too).

What is the best place to ask questions when the package website is not helpful?

Depending what you want to ask, you may use these tools (I keep track of the first three more frequently):
issues page: bug reports and feature requests
knitr mailing list or R-help list: general questions and feedback via email in public; you can also email me in private (see packageDescription('knitr', fields = 'Maintainer'))
StackOverflow: general questions (more public than emails)
Twitter (@xieyihui) or Google+ (+Yihui Xie)

Oh, the tons of arguments like Sweave(..., prefix.string=abc, keep.source=FALSE, foo=bar) are really flexible; why knit() only has so few arguments?

Because I believe putting these arguments in knit() breaks the principle of reproducibility, and so does using environmental variables (see #19 for details).

I love RStudio and Sweave is sweet there; is knitr going to work with RStudio?

I love RStudio too, and yes, knitr is supported in RStudio thanks to RStudio developers since the version 0.96. By the way, I also recommend LyX if you want to use Sweave or knitr.

It is an intelligent wrapper for LaTeX; see http://www.lyx.org for details. I would like to define it as a software package that can both increase the productivity of an experienced LaTeX user by 300%, and decrease it by 500% for a LaTeX novice. Don't use it simply because its GUI is so tempting; it is not MS Word. I have added support for knitr in LyX; see the lyx demo page.

Where are those prompt characters > and +? I feel uncomfortable reading R output without them.

They are removed by default, because I believe they make no sense. This is the reason why I dislike books on R which used > and +; they twist my mind and make my eyes bleed when I read the R code in the books. For those who really want to read R code like > 1+1 instead of 1 + 1, you have the chunk option prompt.

What is the working directory? Can I change my working directory in my code chunks?

You'd better not do this. Your working directory is always getwd() (all output files will be written here), but the code chunks are evaluated under the directory where your input document comes from. Changing working directories while running R code is a bad practice in general. See #38 for a discussion. You should also try to avoid absolute directories whenever possible (use relative directories instead), because it makes things less reproducible.

No, it is not because the box is too narrow (the box uses the current line width); it is because your output is too wide. Use a smaller width option to avoid text output exceeding the page margin, e.g. options(width = 60) (see #44), and tune the option out.width to make your graphics output stay within the margin (see #171).

kframe is a LaTeX environment that I defined for the chunk output, and it does not work well with certain other LaTeX environments, e.g. the center environment; when you want to center your plots, you should use chunk option fig.align='center' instead of putting the whole chunk inside \begin{center} and \end{center}. For more info, see demo page for framed.

see issue #110: you can destroy \Sexpr by %\%Sexpr{code} or comment out R code like \Sexpr{#code}, or just comment out the whole paragraph with % in the beginning of lines.

I have done something cool with knitr; could you add a link in your website?

Many things, e.g. donate me zillions of money (well, I'm kidding), buy me a book from my Amazon wishlist, tweet my links, mention knitr on Google+ or Facebook, or fork this repository and contribute code, or just say hello to me somewhere

Provide feedback