rap: Map over columns of a data frame simultaneously

In romainfrancois/rap: rowwise operations on a data frame

Description

Map over columns of a data frame simultaneously

Usage

wap(.tbl, .f)

lap(.tbl, .f)

rap(.tbl, ...)

Arguments

.tbl	A data frame
.f	a single formula
...	formulas The rhs of each formula uses columns of .tbl, and each stands for a single observation. The lhs of each formula indicates the type, in the vctrs::vec_c() sense. empty or list(): no check is performed on the results of the rhs expression and a list is returned. data.frame(): to indicate that the rhs should evaluate to a data frame of 1 row. The data frames don't need to be of a specific types and are are combined with vctrs::vec_rbind(). A data frame of a specific type, e.g. data.frame(x = integer(), y = double()) The rhs should evaluate to a data frame of that type with 1 row. Any other ptype that makes sense for vctrs::vec_c(). Each result must validate vctrs::vec_size(.) == 1L and are combined with vctrs::vec_c(!!!, .ptype = .ptype) In rap() if the formula is named, the result becomes a new column of the tbl, otherwise the formula is only used for side effects.

Value

• wap() returns a vector of the type specified by the lhs of the formula. The vector validates vec_size() == nrow(.tbl). This is similar to purrr::pmap()

• rap() adds a column to .tbl per formula in ...

Examples

library(purrr)
library(dplyr)
library(tibble)

tbl <- tibble(cyl_threshold = c(4, 6, 8), mpg_threshold = c(30, 25, 20))

# ----- wap
# returns a list of 3 elements
tbl %>%
  wap(       ~ filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold))

# same, i.e. list() is equivalent to empty
tbl %>%
  wap(list() ~ filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold))

# can specify the output type with the formula lhs
tbl %>%
  wap(integer() ~ nrow(filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold)))

# to make data frames
starwars %>%
  wap(data.frame() ~ data.frame(species = length(species), films = length(films)))

# ----- rap: add columns
tbl %>%
  rap(
     x =           ~ filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold),
     n = integer() ~ nrow(x)
  )

# rap is especially useful for iterating over multiple models
starwars %>%
  group_nest(gender) %>%
  rap(
    model =          ~ lm(height ~ mass + birth_year, data = data),
    perf  = double() ~ summary(model)$adj.r.squared
  )

romainfrancois/rap documentation built on June 8, 2019, 5:42 p.m.