https://github.com/romainfrancois/rap

yet another experimental way of processing a data.frame rowwisely
https://github.com/romainfrancois/rap

Last synced: about 1 month ago
JSON representation

yet another experimental way of processing a data.frame rowwisely

• Host: GitHub
• URL: https://github.com/romainfrancois/rap
• Owner: romainfrancois
• License: other
• Created: 2018-11-24T12:08:25.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2019-06-03T10:49:50.000Z (over 5 years ago)
• Last Synced: 2024-07-31T19:25:01.698Z (4 months ago)
• Language: R
• Size: 83 KB
• Stars: 67
• Watchers: 2
• Forks: 7
• Open Issues: 1
• Metadata Files:
  • Readme: README.Rmd
  • License: LICENSE

Awesome Lists containing this project

• jimsghstars - romainfrancois/rap - yet another experimental way of processing a data.frame rowwisely (R)

README

        
---

output: github_document

---

```{r setup, include = FALSE}

knitr::opts_chunk$set(

  collapse = TRUE,

  comment = "#>",

  fig.path = "man/figures/README-",

  out.width = "100%"

)

```

# rap 

[![Lifecycle Status](https://img.shields.io/badge/lifecycle-experimental-blue.svg)](https://www.tidyverse.org/lifecycle/)

[![Travis build status](https://travis-ci.org/romainfrancois/rap.svg?branch=master)](https://travis-ci.org/romainfrancois/rap)

![](https://media.giphy.com/media/l41Yy7rv1mVZNQCT6/giphy.gif)

Experimenting with yet another way to do rowwise operations. 

## Installation

You can install `rap` from gitub

``` r

# install.packages("devtools")

devtools::install_github("romainfrancois/rap")

```

## Why

This offers `rap()` as an alternative to some versions of:

 - `rowwise()` + `do()`

 - `mutate()` + `pmap()`

 - maybe `purrrlyr` ?

 - probably other approaches

`rap()` works with lambdas supplied as formulas, similar to `purrr::map()` 

but instead of `.x`, `.y`, `..1`, `..2`, ...the lambda can use the column names, 

which stand for a single element of the associated vector, in the `[[` sense. 

## rap 

```{r}

library(tidyverse)

library(rap)

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

tbl

tbl %>% 

  rap(x = ~filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold))

```

If the lhs of the formula is empty, `rap()` adds a list column. Otherwise the lhs 

can be used to specify the type: 

```{r}

tbl %>% 

  rap(

    x =           ~ filter(mtcars, cyl == cyl_threshold, mpg < mpg_threshold), 

    n = integer() ~ nrow(x)

  )

```

this example is based on this [issue](https://github.com/tidyverse/purrr/issues/280), 

which has equivalent with `pmap`: 

```{r}

tbl %>%

  mutate(

    x = pmap(

      .l = list(cyl_threshold, mpg_threshold),

      function(cc, mm) filter(mtcars, cyl == cc, mpg < mm)

    ), 

    n = map_int(x, nrow)

  )

```

## wap

```{r}

library(dplyr)

starwars <- head(starwars)

# creates a list of length 1 integer vectors

# because type not specified

starwars %>% 

  wap(~length(films)) 

# using the lhs to specify the type

starwars %>% 

  wap(integer() ~ length(films))

# list of data frames

starwars %>% 

  wap(~ data.frame(vehicles = length(vehicles), starships = length(starships)))

# Specify type as data.frame() row binds them

starwars %>% 

  wap(data.frame() ~ data.frame(vehicles = length(vehicles), starships = length(starships)))

```

## zest_join

`r emo::ji("lemon")` `zest_join()` is similar to `dplyr::nest_join()` but 

you control what goes in the nested column. `Z` is `N` but `r emo::ji("arrow_heading_down")`. 

```{r}

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

tbl %>%

  zest_join(mtcars, data = ~cyl == cyl_threshold & mpg < mpg_threshold)

```

In the rhs of the formula : 

  - `cyl` and `mpg` refer to columns of `mtcars`

  - `cyl_threshold` and `mpg_threshold` refer to the current value from `tbl` because these columns don't exist in mtcars. If you wanted to refer to columns that are present both in mtcars and tbl you would have to unquote the columns in tbl with the unquoting operator, e.g. !!cyl