Stephen Royle

I like to set up a standardised directory structure for RStudio
projects. The idea came from
[here](https://martinctc.github.io/blog/rstudio-projects-and-working-directories-a-beginner's-guide/).
In brief, the structure is:

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .plain; .title: .; .notranslate title=""}
Data/
Output/Data/
Output/Plots/
Script/
```
:::

My typical workflow is therefore to:

- select File \> New Project in RStudio
- make a new directory and RProj file
- then use this [R script or these shell
  commands](https://gist.github.com/quantixed/42625f988a7b5da25b7e333c4a660b97)
  to setup the directories.

So far, so good. However, this process is a bit tedious.

I find it irritating to open up RStudio (which opens up a previous
project by default) and then go through these steps. Recently, I had to
make several R projects for a paper and I thought there must be a
quicker way. Here's my solution.

## The quick way: shell function {#the-quick-way-shell-function .wp-block-heading}

I made a shell function to set up an R Project with the correct
directory tree. The resulting `.Rproj` file can just be double-clicked
and you are good to go!

I previously wrote about [shell aliases and
functions](https://quantixed.org/2023/01/10/over-and-over-quick-ways-to-reuse-handy-shell-one-liners/){data-type="post"
data-id="2966"}. I am using zsh on macOS, but this should all work on
bash and possibly shells.

As described before, I had to initialise a file for zsh functions.

In the home directory (`~`), open `.zshrc` with your favourite editor
e.g.

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .bash; .title: .; .notranslate title=""}
nano .zshrc
```
:::

Then add the following to the file:

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .plain; .title: .; .notranslate title=""}
# Alias definitions
if [ -f ~/.zsh_functions ]; then
    . ~/.zsh_functions
fi
```
:::

Now make and edit `zsh_functions`

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .bash; .title: .; .notranslate title=""}
nano .zsh_functions
```
:::

Now add the function:

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .plain; .title: .; .notranslate title=""}
function setup_rproj() {
    if [[ $# -eq 0 ]] ; then
        echo 'Name of new directory required'
        return 0
    fi

    # setup preferred directory structure
    mkdir "$1"
    mkdir "$1"/Data;\
    mkdir "$1"/Output;\
    mkdir "$1"/Output/Data;\
    mkdir "$1"/Output/Plots;\
    mkdir "$1"/Script

    /bin/cat <<EOM >"$1"/"$1".Rproj
Version: 1.0

RestoreWorkspace: Default
SaveWorkspace: Default
AlwaysSaveHistory: Default

EnableCodeIndexing: Yes
UseSpacesForTab: Yes 
NumSpacesForTab: 2
Encoding: UTF-8

RnwWeave: Sweave
LaTeX: pdfLaTeX
EOM

    # end of function here
}
```
:::

And save by ctrl + x and y.

## How it works {#how-it-works .wp-block-heading}

The idea is that the user navigates to where they want to make the R
Project, i.e. `cd` to the location. Then call the function like this:

::: {.wp-block-syntaxhighlighter-code}
``` {.brush: .plain; .title: .; .notranslate title=""}
setup_rproj analysis123
```
:::

The script will make a folder in the current directory called
analysis123, in there it will place the directory structure described
above, and then make the `.Rproj` file.

If you use a different `.Rproj` file the function needs to be edited
accordingly. Finally, if the user calls `setup_rproj` without specifying
a project name, the function will return an error message.

If the directory already exists, nothing will be overwritten.

---

Part of an occasional series of tech tips.


I like to set up a standardised directory structure for RStudio projects. The idea came from here. In brief, the structure is: Data/ Output/Data/ Output/Plots/ Script/  My typical workflow is therefore to: select File &gt; New Project in RStudio make a new directory and RProj file then use this R script or these shell commands to setup the directories. So far, so good. However, this process is a bit tedious.

Messaggi di Rogue Scholar

Tips From The Blog XIX: initialising a new RStudio project