Pipes and functions in R

Learning objectives

• Demonstrate how to use the pipe
• Demonstrate piping alternatives
• Define functions and their purpose
• Analyze a user-written function and explain how it works
• Demonstrate conditional execution
• Practice writing functions

Notes

April 27, 2020

Pipes in R

• Think back to the diamonds dataset
• filter(), group(), and then summarize()

Approaches to Coding

Intermediate Steps

• diamonds_# <- function
• Unclear because the names are not intuitive
  • Might forget what the name of the object means
  • Make spelling mistakes
  • Might start running out of memory (not as common)
• This works, but there are a couple of concerns when writing code this way all the time

Overwrite the original

diamonds_t <- diamonds
diamonds_t <- filter()
...

• If something goes wrong, you basically have to start over adn retrieve the original dataset

Function Composition

summarize(
  group_by(
    filter(diamonds, color == "I"),
    cut
  ),
  price = mean(price)
)

• Now we're writing code in sort of a backwards manner
• Again, not very intuitive

All of these approaches produce the exact same output, but we want to make sure that our code is interpretable and reproducible_

Piping

• A lot more intuitive
• Implicit
• Remember: You don't assign <- something to an object inside the pipe operator
• Depending on complexicity of task, you may have a lot of steps
  • Break piped operation into a smaller set of tasks and save output
• Next line of code is automaticaly indented when you are using a piped operator
• For all of the tidyverse functions we use the %>% at the end of a line.
  • ggplot2 is different in that it uses a plus symbole +

The %>% comes from the maggritr package. There are other piping operators in the package, including the "tee" pipe %T% and %<>%

Functions in R

• Easy to use
• Self-documenting
• Easy-ier to debug
• If you have copied and pasted a block of code more than twice, convert it to a function

Rescale Function

• Name
• Arguments
• Body

How to use a function

• When using functions, by default the returned object is merely printed to the screen

Conditional execution

if (condition) {
  # code executed when condition is TRUE
} else {
  # code executed when condition is FALSE
}

• Recoding variables is where you would use if() statements
• Need a vectorized statement check for multiple conditions
  • if_else()

FIZZBUZZ