README.md

In dirkschumacher/defmacroex: An Example of defmacro

defmacroex

Just an example of how to use the defmacro package.

Example

The package exports a function called analyze_dataset with the following code that defines its own pipe function and a zero overhead logging macro.

analyze_dataset <- function(data) {
  log("Computing result")
  assert("is a data frame", stopifnot(is.data.frame(data)))
  assert("has right columns", stopifnot(c("hp", "cyl") %in% colnames(data)))
  result <- data %>%
    dplyr::filter(hp > constexpr(50 + 50 + qnorm(0.975))) %>%
    dplyr::group_by(cyl) %>%
    dplyr::summarise(dplyr::n())
  log("Returning result")
  result
}

During package load, the code gets replaced by the expanded version of the macro. This means the function does not have any runtime overhead.

defmacroex::analyze_dataset
#> function (data) 
#> {
#>     result <- dplyr::summarise(dplyr::group_by(dplyr::filter(data, 
#>         hp > 101.95996398454), cyl), dplyr::n())
#>     result
#> }
#> <bytecode: 0x7fc16f83fe70>
#> <environment: namespace:defmacroex>

Had we enabled logging through an environment variable, the logging code would have been added to the function:

unloadNamespace("defmacroex")
withr::with_envvar(c("LOGGING" = "1", "ASSERT" = "1"), {
  defmacroex::analyze_dataset
})
#> function (data) 
#> {
#>     message("Computing result")
#>     stopifnot(is.data.frame(data))
#>     stopifnot(c("hp", "cyl") %in% colnames(data))
#>     result <- dplyr::summarise(dplyr::group_by(dplyr::filter(data, 
#>         hp > 101.95996398454), cyl), dplyr::n())
#>     message("Returning result")
#>     result
#> }
#> <bytecode: 0x7fc16eaa42f8>
#> <environment: namespace:defmacroex>

dirkschumacher/defmacroex documentation built on Dec. 3, 2019, 6:59 p.m.