r-statistics.co by Selva Prabhakaran

About

Infix Functions in R: Write Your Own %op% Operators

An infix function sits between its two arguments: x + y instead of `+`(x, y). R lets you create custom infix operators using the %name% syntax.

You use infix operators constantly: +, -, *, %>%, %in%, <-. What most R users don't know is that you can define your own. Any function named %something% automatically works as an infix operator.

Built-in Infix Operators

Every operator in R is actually a function. You can call them in prefix form.

# Normal (infix) usage cat("3 + 4 =", 3 + 4, "\n") cat("10 %% 3 =", 10 %% 3, "\n") # modulo cat("2 %in% c(1,2,3) =", 2 %in% c(1,2,3), "\n") # Prefix form (function call syntax) cat("`+`(3, 4) =", `+`(3, 4), "\n") cat("`%%`(10, 3) =", `%%`(10, 3), "\n") cat("`%in%`(2, c(1,2,3)) =", `%in%`(2, c(1,2,3)), "\n")


  






Creating Custom Infix Operators



Define a function with a name wrapped in %...% and backticks.





  


    
# Null coalescing operator (like Python's 'or' or JS's '??')
`%||%` <- function(a, b) {
  if (is.null(a)) b else a
}

x <- NULL
y <- 42
cat("x %||% y:", x %||% y, "\n")  # 42
cat("10 %||% y:", 10 %||% y, "\n") # 10


    


      

      

    


    


  








  


    
# String concatenation operator
`%+%` <- function(a, b) paste0(a, b)

greeting <- "Hello" %+% " " %+% "World" %+% "!"
cat(greeting, "\n")

# Between operator
`%between%` <- function(x, range) {
  x >= range[1] & x <= range[2]
}

values <- c(3, 7, 12, 1, 9, 15)
cat("Between 5 and 10:", values %between% c(5, 10), "\n")


    


      

      

    


    


  






Practical Custom Operators



Default Value Operator





  


    
# Use a default when value is NA (not NULL)
`%na%` <- function(x, default) {
  ifelse(is.na(x), default, x)
}

data <- c(1, NA, 3, NA, 5)
cat("Original:", data, "\n")
cat("With default:", data %na% 0, "\n")


    


      

      

    


    


  






Not-In Operator





  


    
# The opposite of %in%
`%notin%` <- function(x, table) {
  !(x %in% table)
}

fruits <- c("apple", "banana", "cherry", "date")
exclude <- c("banana", "date")

cat("Not in exclusion list:", fruits[fruits %notin% exclude], "\n")


    


      

      

    


    


  






Approximate Equality





  


    
# Check if two numbers are approximately equal
`%~%` <- function(a, b) {
  abs(a - b) < .Machine$double.eps^0.5
}

cat("0.1 + 0.2 == 0.3:", 0.1 + 0.2 == 0.3, "\n")  # FALSE!
cat("0.1 + 0.2 %~% 0.3:", (0.1 + 0.2) %~% 0.3, "\n")  # TRUE


    


      

      

    


    


  






How the Pipe Operators Work



The pipe operators %>% and |> are just infix functions (though |> is a language-level construct).





  


    
# The magrittr pipe is essentially:
`%pipe%` <- function(lhs, rhs) {
  rhs(lhs)
}

result <- 1:10 %pipe% mean %pipe% round
cat("Result:", result, "\n")


    


      

      

    


    


  






Rules for Custom Operators







Rule
Example






Must start and end with %
%op%, %+%, %between%




Name goes inside the % signs
%myop%




Must be defined with backticks
` %myop% <- function(a, b) ... `




Always takes exactly 2 arguments
Left operand and right operand




Can contain letters, numbers, .
%do.call%, %v2%







Practice Exercises



Exercise 1: Range Check Operator



Create a %within% operator that checks if values fall within a named range.





  


    
# Create %within% that works like:
# 5 %within% c(1, 10)   -> TRUE
# 15 %within% c(1, 10)  -> FALSE
# c(3, 7, 12) %within% c(5, 10)  -> FALSE TRUE FALSE



    


      

      

    


    


  







Click to reveal solution

```r


  

    
`%within%` <- function(x, range) {
  x > range[1] & x < range[2]  # Exclusive bounds
}

cat("5 %within% c(1,10):", 5 %within% c(1, 10), "\n")
cat("15 %within% c(1,10):", 15 %within% c(1, 10), "\n")

values <- c(3, 7, 12, 5, 10)
cat("Values:", values, "\n")
cat("Within (5,10):", values %within% c(5, 10), "\n")

    

      
      
    

    

  




**Explanation:** Custom operators are normal functions that happen to use infix syntax. The left side becomes the first argument, the right side becomes the second.





Summary







Operator
Purpose
Package






%>%
Forward pipe
magrittr




%in%
Membership test
base




%%
Modulo
base




%*%
Matrix multiply
base




%o%
Outer product
base




`%

%`
Null coalescing
rlang/custom




%+%
ggplot2 layer add
ggplot2







FAQ



Are custom operators slower than regular functions?



No. An infix operator is just a regular function with special syntax. There is no performance penalty.



Can I override built-in operators like + or *?



Yes, but you shouldn't for primitive types. You can define methods for custom S3/S4 classes: "+.myclass" <- function(a, b) .... Overriding for base types creates confusing code.



Why do custom operators need % signs?



The %...% syntax is R's way of distinguishing user-defined infix operators from built-in ones. Without it, the parser wouldn't know if myop is a variable or an operator.



What's Next?




        


        
      


      

        

        
© 2016- Selva Prabhakaran.
          This work is licensed under the Creative Commons License.