README.md

In nayefahmad/tidyverseInPackagesTest: What the Package Does (One Line, Title Case)

tidyverseInPackagesTest

The goal of tidyverseInPackagesTest is to demo the use of tidyverse functions in the development of new packages.

This is an example as part of the "Building tidy tools" workshop at rstudio::conf 2020

Installation

You can install the released version of tidyverseInPackagesTest from GitHub with:

devtools::install_github("tidyverseInPackagesTest")

Example

This is a basic example which shows you how to solve a common problem:

library(tidyverseInPackagesTest)
## basic example code

Notes

• Data masking: means you can refer to data vars ("statistical vars") as if they are env vars ("programming variables"). E.g. "mtcars" is an env var that is always available. "cyl" is a data var associated with mtcars.

• Why tidyverse uses data masking: just easier, faster.

• Disadvantage: programming with tidyverse (i.e. using tidyverse fns in your own fns) is harder.

• If you refer to a data var inside your function, you have to preface it with .data from the rlang package.

• To properly document the use of fns from other packages (e.g. %>% ), prob best to both use usethis::use_package() and use @importFrom magrittr %>% in the roxygen documentation of the specific function.

  • Hadley: @importFrom changes the NAMESPACE - affects all functions in the package. I prefer not to put it in individual functions, but to create a separate file, using usethis::use_package_doc()
  • NAMESPACE is about functions; DOCUMENTATION is about packages

Indirection

See tidy evaluation. Two methods of fixing:

1. Using square brackets

not recommended:

cyl_plot <- function(var){ ggplot2::ggplot(mtcars) + ggplot2::geom_bar(ggplot2::aes(.data[[var]])) + ggplot2::coord_flip()}

2. Using "embracing"

This is basically the latest way of doing quoting/unquoting.

nayefahmad/tidyverseInPackagesTest documentation built on Feb. 4, 2020, 6:20 p.m.