Yes, we all love to use the magical knitr package (Thanks to Yihui) to render .Rmd for reports. And I honestly think knitr, .Rmd, and R Studio makes R programming much better and simpler.

However, here’s a minor defect about knitr, by defult, it will only search the same directory as your .Rmd file to locate the dataset or files you need to pull in. It works fine when you have all your data files and .Rmd in the same folder (which you really shouldn’t). But what would happen if you have accumulated tons of files in that same directory? You get lost.

So you start to organize your scrips and files into subdirectories docs/, data/, script/, and you will need to configure knitr.

You can use setwd() with knitr to change the working directory, but you will have to run setwd('../..') at the start of any chunk that loads data (which is unpractical and ugly). The elegant way to do this is to use the package option root.dir to specify the root directory instead of forcibly using the directory of the input document. e.g. knitr::opts_knit$set(root.dir = ...)

This will set the root directory to your project folder. Knitr will now look for all files from this root folder rather than the folder .Rmd stored in.

Example

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
knitr::opts_knit$set(root.dir = '/Users/frankzheng/Desktop/test/data')
```

```{r}
library('readxl')
```

```{r}
getwd()
read_excel("data.xlsx")
```

Note

• knitr::opts_knit$set(root.dir = ...) works also in normal chunks, not only in the “setup” one.
• The working directory starts to change in the following chunks after the chuck that contains knitr::opts_knit$set(root.dir = ...).