details.md

R details

If-Else Statements

if from base R

• fast, but not vectorial
• must receive only one value, not a vector
• if returns a value which can be assigned
• always leave a space after if

    a <- if (one_logical_condition) value_when_TRUE else value_when_FALSE

• nested ifs require curly brackets

if (one_logical_condition) {
    # do something when condition is TRUE
} else {
    # do something else when condition is FALSE
}

ifelse from base R

• vectorized version
• faster than if when dealing with a vector
  • https://stackoverflow.com/questions/34005423/is-if-faster-than-ifelse
• treat it as a function

dplyr::if_else

• as if_else but can deal with NAs and validates data types and vector lengths

Multi if statements

Nested ifs

• allow for different conditions

if (one_logical_condition) {
    # do something when first condition is TRUE
} else if (second_logical_condition) {
    # do something else when second condition is FALSE (and first is FALSE)
} else {
    # do something when both conditions are FALSE
}

switch from base R

• think of it as multi-case if based on only one
• one expression that evaluates to one number or one string
• good practice to have a catch-all case (first unnamed argument)

purrr::when

• think of it as vectorized if, with multiple conditions
• uses formulas which must evaluate to a single value
• good practice to have a catch-all case

dplyr::case_when

• think of it as multi-case if_else
• uses formulas which must evaluate to a vector(!)
• good practice to have a catch-all case

Data Frames

Row names

Avoid row names in data frames

• Many functions in tidyverse remove them
• Not compatible with database tables
• other issues

Functions

Arguments

• use match.arg() to validate function arguments