In AshirBorah/cp_bootcamp_r_tutorials:

knitr::opts_chunk$set(echo = TRUE)
library(learnr)
tutorial_options(exercise.timelimit = 10, exercise.blanks = "___+")

What are functions?

• Self-contained modules of code that accomplish a specific task

output <- do_the_thing(input1, input2)

• Inputs (arguments) can include:

• the objects on which the function acts

• additional parameters that specify how the function acts (options)

Example functions

Calculating mean of a vector

num_vec <- 1:10
mean(num_vec)

---

Functions don't need to provide output, or take inputs

• Save data to a file (no output)

save(object_to_save, file = 'data_file')

• Get 'working directory' (no input needed)

getwd()

Built-in functions

Mathematical functions

Functions that operate on individual values:

• sqrt()
• abs()
• log()
• exp()

Mathematical functions

Functions that estimate statistics across a set of values:

• mean()
• median()
• max()

x <- c(1, 1, 2, 3, 5, 8, 13)

Getting help

• Understanding how to get help effectively is the most important skill!

?my_function
help(my_function)

• Or use the Help tab in RStudio to search.

Example help documentation

Tips on using functions

• Use tab-complete to quickly check function inputs.

• Best to name arguments explicitly (at least the optional ones)

round(pi, digits = 2) 

See examples of function being used!

example('round')

The 'combine' function

What are the inputs and outputs?

glengths <- c(4.6, 3000, 580)

Creating functions (the general pattern)

name_of_function <- function(argument1, argument2) {
    *code that does something based on inputs*
    return(some_results)
}

• function is a special keyword for creating functions

• Assign the function to a variable (give it a name)

• List inputs ('arguments')

• Use curly braces to wrap text that performs some action on any inputs

• (Optionally) return a value from the function

Creating functions (a simple example)

combine_two_numbers <- function(num1, num2) {
  result <- num1 + num2
  return(result)
}

combine_two_numbers(1, 2)

Don't forget to return a result

Making 'optional' inputs

Assign a default value to inputs with = to make them optional

subtract_from_vec <- function(data_vec, subtract_val = 1) {
  new_data_vec <- data_vec - subtract_val
  return(new_data_vec)
}
test_vec <- 1:10
subtract_from_vec(test_vec)
subtract_from_vec(test_vec, subtract_val = 2)

What are packages?

• Collections of functions, and data, designed to add specific functionality beyond 'base R'

• Need to install before you can use (only do this once)

install.packages('tidyverse')

• Load the package into your R session (do this every session)

library(tidyverse)

Getting help with packages

• Use the help function

help(package = 'lubridate')

• Use the 'Packages` tab in Rstudio

Quick note on Bioconductor

• CRAN is the default 'app store' for R packages

• Bioconductor is another that has many useful bioinformatics related packages

install.packages("BiocManager")
BiocManager::install("GenomicRanges")

• (Looks weird, but just copy-paste)

• Once installed they're just like other packages

library(GenomicRanges)

Word of warning about function names

• Multiple packages can have the same names for different functions.

• Which one R thinks you're using in a given instance can be hard to know.

• Cause of many painful 'bugs'

• When in doubt it's best to specify which package the function comes from:

library(lubridate)
library(here)
lubridate::here()
here::here()

Getting more help

• Pay attention to error messages and warning messages

• Using Google

• Search error messages verbatim (but remove stuff specific to your usage).

• Add "r" "error"

• Stackoverflow: almost every problem you encounter is answered here (often 10 different ways!)

Google -> StackOverflow

Key concepts recap

Using functions:

• output <- do_a_thing(input_obj, param1 = my_val)

Writing your own functions:

do_a_thing <- function(input_obj, param1) {
  **STUFF**
  return(result)
}

• Getting help with functions: help(my_fun).

• install.packages() and library() to install/use packages

• Google (and stackoverflow, Biostars, etc.) are your friends!

Now for hands-on practice!

• I'll open up breakout rooms

• Go to broad.io/cp_r_bootcamp

• Work through Chapters 1 and 2 Practice

• There are clickable 'hints' throughout.

• Use slack for Q&A.

• Use 'ask for help' button and TA can jump into Breakout room to help

• Group Slack channels and 'huddle' audio spaces might be helpful too.

Tips for using web site

• Your web browser will save your progress!

• You can reset your work in a single code chunk with the 'start over' button

• Your work in each code chunk is independent (results don't carry over)

• You can reset the whole notebook with the 'Start over' button on the left panel (careful!)

AshirBorah/cp_bootcamp_r_tutorials documentation built on May 16, 2024, 3:24 p.m.