functions with R

Presentation materials: https://github.com/ecodatasci-tlv/functions

Sometimes, we tend to repeat ourselves when coding: Repeating similar analyses, getting data ready for plots, etc. Functions have multiple advantages over copy and pasting chuncks of code:

• Naming a function improve readability of code (“write codes for humans”)
• Changing paramters and conditions in one place.
• Eliminate chance of mistakes due to copy-pasting.

A good rule of thumb is to not copy and paste code more than twice.

df$a <- (df$a - min(df$a, na.rm = TRUE)) / 
  (max(df$a, na.rm = TRUE) - min(df$a, na.rm = TRUE))
df$b <- (df$b - min(df$b, na.rm = TRUE)) / 
  (max(df$b, na.rm = TRUE) - min(df$a, na.rm = TRUE))
df$c <- (df$c - min(df$c, na.rm = TRUE)) / 
  (max(df$c, na.rm = TRUE) - min(df$c, na.rm = TRUE))
df$d <- (df$d - min(df$d, na.rm = TRUE)) / 
  (max(df$d, na.rm = TRUE) - min(df$d, na.rm = TRUE))

Suprisingly, not so much. Focus on your function doing just one thing.

Most of R's functions are less than 12 lines long!

Function names should be kept short yet informative. Remember: functions are meant to help codes be reusable and readable.

What are good names for these two functions?

f1 <- function(string, prefix) {
  substr(string, 1, nchar(prefix)) == prefix
}
f2 <- function(x) {
  if (length(x) <= 1) return(NULL)
  x[-length(x)]
}

Sometimes, you won't save the function like seen below, but rather - use it directly in a code:

matrix_of_numbers <- matrix(1:100,10,10)
apply(matrix_of_numbers,2,function(x) x^2)

      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10]
 [1,]    1  121  441  961 1681 2601 3721 5041 6561  8281
 [2,]    4  144  484 1024 1764 2704 3844 5184 6724  8464
 [3,]    9  169  529 1089 1849 2809 3969 5329 6889  8649
 [4,]   16  196  576 1156 1936 2916 4096 5476 7056  8836
 [5,]   25  225  625 1225 2025 3025 4225 5625 7225  9025
 [6,]   36  256  676 1296 2116 3136 4356 5776 7396  9216
 [7,]   49  289  729 1369 2209 3249 4489 5929 7569  9409
 [8,]   64  324  784 1444 2304 3364 4624 6084 7744  9604
 [9,]   81  361  841 1521 2401 3481 4761 6241 7921  9801
[10,]  100  400  900 1600 2500 3600 4900 6400 8100 10000

To further streamline your code, use iterations to repeat chuncks of code. Most basic are for and while loops. However, there are more ways to iterate code.

Iterations have three main components: an output, a sequence to iterate over, and the body of code.

library(tictoc)
n_times <- 50000
tic()
a <- NULL
for(i in seq_len(n_times)){
    a <- c(a, i^2)
    a
}
toc()
tic()
a <- vector("double", n_times)
for(i in seq_len(n_times)){
    a[i] <- i^2
}
toc()

for (i in (1:10)) is a sequence. So is while TRUE. for (i in seq_along(vector)) is a better way of sequencing if you might get a vector of length 0 like so:

y <- vector("double", 0)
seq_along(y)

integer(0)

#> integer(0)
1:length(y)

[1] 1 0

#> [1] 1 0

That's a common phrase when working with data. So common that there is a tidyverse package that does it for you.

We will discuss it shortly…

That means you can wrap loops within a function and just call that function when neccessary. Remember - limit your copy and paste as much as possible!

library(tidyverse)
data <- tibble(a=rnorm(10),
               b=rnorm(10))
col_means <- function(dataframe){
  output <- vector("double",ncol(dataframe))
  for (i in seq_along(dataframe)){
    output[[i]] <- mean(dataframe[[i]])
  }
  return(output)
}
col_means(data)

[1] -0.2508394 -0.3507583

Create a function that returns the full song for any number of any vessel (bottles,cans, even boxes… ) of any drink (But no Jägermeister please):

99 bottles of beer on the wall, 99 bottles of beer. Take one down, pass it around - 98 bottles of beer on the wall