knitr_in {drake}R Documentation

Declare knitr/rmarkdown source files as dependencies. [Stable]


knitr_in() marks individual knitr/R Markdown reports as dependencies. In drake, these reports are pieces of the pipeline. R Markdown is a great tool for displaying precomputed results, but not for running a large workflow from end to end. These reports should do as little computation as possible.





Character strings. File paths of knitr/rmarkdown source files supplied to a command in your workflow plan data frame.


Unlike file_in() and file_out(), knitr_in() does not work with entire directories.


A character vector of declared input file paths.


drake_plan() understands special keyword functions for your commands. With the exception of target(), each one is a proper function with its own help file.

See Also

file_in(), file_out(), ignore(), no_deps()


## Not run: 
isolate_example("contain side effects", {
if (requireNamespace("knitr", quietly = TRUE)) {
# `knitr_in()` is like `file_in()`
# except that it analyzes active code chunks in your `knitr`
# source file and detects non-file dependencies.
# That way, updates to the right dependencies trigger rebuilds
# in your report.
# The mtcars example (`drake_example("mtcars")`)
# already has a demonstration


# Now how did drake magically know that
# `small`, `large`, and `coef_regression2_small` were
# dependencies of the output file ``?
# because the command in the workflow plan had
# `knitr_in("report.Rmd")` in it, so drake knew
# to analyze the active code chunks. There, it spotted
# where `small`, `large`, and `coef_regression2_small`
# were read from the cache using calls to `loadd()` and `readd()`.

## End(Not run)

[Package drake version 7.13.9 Index]