Writing functions

Your Turn 1

Re-write this ggplot2 theme as a function. Call it theme_avalanche_h().

Run this code to test that your function works

Your Turn 1

theme_avalanche_h <- function() {
  theme_minimal(base_size = 14) +
    theme(panel.grid.minor = element_blank(), panel.grid.major.y = element_blank())
} 
residents_per_sector <- 
  data.frame(
    sector = as.factor(1:8),
    residents = c(1000, 2034, 4594, 2304, 8093, 1200, 300, 2398)
  )
ggplot(residents_per_sector, aes(forcats::fct_reorder(sector, residents), residents)) +
  geom_col() + 
  coord_flip() + 
  xlab("sector") + 
  theme_avalanche_h()

use_r("file_name")

Write a new file to R/

devtools: loading vs. building

load_all(): fast, all functions available

build(): builds and installs the package

devtools: loading vs. building

load_all(): fast, all functions available

build(): builds and installs the package

Keyboard shortcuts

load_all(): CMD/CTRL + Shift + L

build(): CMD/CTRL + Shift + B

Your Turn 2

Create a new file with use_r() called "db_con"

Put this function in the file and save it

Use load_all() to load the package function.

Run this code to make sure it works:

Your Turn 2

use_r("db_con")

in R/db_con.R:

db_con <- function(dbname = "residents_per_sector") {
  dbname <- match.arg(dbname)
  # We will set up real database connections later. For now, 
  # we'll just return some hard-coded data instead.
  data.frame(
    sector = as.factor(1:8),
    residents = c(1000, 2034, 4594, 2304, 8093, 1200, 300, 2398)
  )
}

Your Turn 2

Testing for suggested packages

if (requireNamespace("shiny", quietly = TRUE)) {
  # ... shiny code
}

Your Turn 3

Fix the code in R/themes.R to use ggplot2:: instead of library(ggplot2)

Run use_package("ggplot") to add ggplot2 to Imports

Re-load the package (Cmd/Ctrl+Shift+L) and run this code to make sure it works

Your Turn 3

use_package("ggplot2")

in R/themes.R:

theme_avalanche <- function() {
  ggplot2::theme_minimal(base_size = 14) +
    ggplot2::theme(panel.grid.minor = ggplot2::element_blank())
}

Your Turn 3

Packages so good they get their own functions

Packages so good they get their own functions

use_tibble()

use_data_table()

use_pipe()

use_rcpp*()

Your Turn 4

We need roxygen2 for this exercise. We'll learn more about it in the next module. For now, just run use_roxygen_md()

Run use_tibble() and use_data_table()

In R/get_data.R, edit the function to be able to return a data table: Add the argument data_table = FALSE. If data_table is TRUE, convert the data frame with data.table::as.data.table()

Run this code to make sure it works

Your Turn 4

use_roxygen_md()
use_tibble() 
use_data_table()

get_resident_data <- function(data_table = FALSE) {
  residents_per_sector <- db_con("residents_per_sector")
  if (data_table)
    return(data.table::as.data.table(residents_per_sector))
  tibble::as_tibble(residents_per_sector)
}

Your Turn 4

##    sector residents
## 1:      1      1000
## 2:      2      2034
## 3:      3      4594
## 4:      4      2304
## 5:      5      8093
## 6:      6      1200
## 7:      7       300
## 8:      8      2398

data.table::is.data.table(res_data)

## [1] TRUE

Your Turn 4: Stretch goal

Run use_pipe() to add the magrittr pipe to your package. What changed?

Organizing .R files

Organizing .R files

Less than 1:1, more than all:1

Organizing .R files

Less than 1:1, more than all:1

utils.R

Organizing .R files

Less than 1:1, more than all:1

utils.R

zzz.R, .onLoad

See R Packages, ed. 2 for more.