Relevant Quotes

“to err is human; to forgive, divine” - Alexander Pope, “Essay on Criticism”

“to err is human; to really screw up, you need a computer” - Internet

“Programs must be written for people to read, and only incidentally for machines to execute.” ― Harold Abelson, Structure and Interpretation of Computer Programs

Presentation overview

• Overview and intended audience
• Introduction
  • Humanizing mistakes
  • Feelings exercise
  • The B.R.I.E Method 🧀
• Common sources of error messages
  • Data types & structures
  • Data wrangling
• Additional Learning Resources
• Questions & Discussion

Intended audience

• This presentation is intended for beginners to the R language and environment
• Those who want to overcome fears of decoding messages
• Those who want to build portable skills to asses their code and minimize downtime due to R’s sometimes confusing & cryptic errors.

Introduction

Everyone makes mistakes

• Errors (and their messages) happen to everyone, even experienced programmers.
• Learning how to solve and interpret errors will help us to write efficient and elegant code.
• Developing skills to cope with the frustration of error message will make for a more enjoyable time crafting programs.

All the feels! | Describe your feelings about this picture

B.R.I.E Method 🧀

• First steps in confidently approaching error messages is by following the B.R.I.E Method
  • Breathe and relax your mind
  • Re-read the error message and focus on key wording
  • Isolate the Error within your code to reduce complexity
• Hopefully this method can act as a guide to approaching and solving errors in your R code (and other programming languages as well!)

Common Sources of Error Messages

Data types & structures

• Many error messages start from applying the wrong method or function with a data structures.
• R base has 5 common data structures:
  • Atomic Vector: logical, integer, double (often called numeric), and character.
  • Matrix
  • Array
  • List
  • Data Frame

Error #1: non-numeric argument to binary operator

name_string = "Jasmine"
name_string * 2

## Error in name_string * 2: non-numeric argument to binary operator

• To decode this error the key statement is ‘binary operator’. Multiplication requires two numeric values around the * symbol.
• This error informs the user that one of the values supplied to the multiplication recipe is not a number!
• Note: A unary operator is a + (positive) or - (negative) in-front of a number.

Error #2: invalid ‘type’ (character) of argument

vec <- c(1, 2, 3, 4, 5, "hello world")
sum(vec)

## Error in sum(vec): invalid 'type' (character) of argument

• This error states that the character ‘type’ cannot be included in the aggregate function sum(). Only numeric or integer values can be summed.
• Once again this is a improper mixing of data types numeric and characters.

Data wrangling

• Data wrangling and cleaning can take the majority of time in the data science workflow.
• Many type of errors can arise when subsetting and manipulating multi-dimensional data objects.
• With different types of packages it can be common to get flustered by confusing errors that can be function specific such as requiring an input argument to be a list instead of a data frame

Error #3: all arguments must have the same length

df <- data.frame("age" = c(32, 31, 34), 
                 "pet" = c(F, T, T), 
                 "name" = c("Adam", "Blake", "Anders"))
table(df$age, df$pets)

## Error in table(df$age, df$pets): all arguments must have the same length

• The error describes that the two supplied arguments to table() are not the same length which is required for cross tabulation. The spelling mistake (pet versus pets) actually has a length of 0 which is a mismatch from df$age which has a length 3 but the error message does not indicate that the column name was misspelled.

Additional Learning Resources

• Advanced R by Hadley Wickham http://adv-r.had.co.nz/
• R-bloggers post on weird R errors http://www.r-bloggers.com/translating-weird-r-errors/

Questions & Discussion!