R support¶
Ploomber officially supports R. The same concepts that apply to Python
scripts apply to R scripts; this implies that R scripts can render as notebooks
in Jupyter and the cell injection works. The only difference is how
to declare upstream dependencies:
For the R Markdown format (.Rmd):
```{r, tags=c("parameters")}
upstream = list('one_task', 'another_task')
```
If you prefer, you can also use plain R scripts:
# %% tags=["parameters"]
upstream = list('one_task', 'another_task')
#
If your script doesn’t have dependencies: upstream = NULL
To read more about how Ploomber executes scripts and integrates with Jupyter, check the Jupyter Integration guide.
Configuring R environment¶
To run R scripts as Jupyter notebooks, you need to install Jupyter first, have an existing R installation and install the IRkernel package.
If you are using conda and a environment.yml file to manage
dependencies, keep on reading. Otherwise, read the IRkernel installation
instructions.
Setting up R and IRkernel via conda¶
Even if you already have R installed, it is good to isolate your
environments from one project to another. conda can install R inside your
project’s environment.
Add the following lines to your environment.yaml:
name: some_project
dependencies:
# ...
# existing conda dependencies...
- r-base
- r-irkernel
# optionally add r-essentials to install commonly used R packages
- pip:
# ...
# existing pip dependencies...
- ploomber
For more information on installing R via conda
click here.
Once you update your environment.yml, re-create or update your environment.
Finally, activate the R kernel for Jupyter. If you’re using Linux or macOS:
echo "IRkernel::installspec()" | Rscript -
If using Windows, start an R session and run IRkernel::installspec() on it.
Interactive example¶
Click the button above to see an interactive example (no installation needed, but takes about a minute to be ready):