Module 11: Debugging R Functions

For this assignment, I worked on debugging a function that identifies outlier rows in a numeric matrix using the Tukey rule. The original code had a bug that I had to find and fix.

The Original Buggy Code

The function was supposed to flag rows where every column contains an outlier. Here's what I started with:

tukey.outlier = function(x) {

  Q1 = quantile(x, 0.25, na.rm = TRUE)

  Q3 = quantile(x, 0.75, na.rm = TRUE)

  IQR = Q3 - Q1

  lower = Q1 - 1.5 * IQR

  upper = Q3 + 1.5 * IQR

  return(x < lower | x > upper)

}


tukey_multiple = function(x) {

  outliers = array(TRUE, dim = dim(x))

  for (j in 1:ncol(x)) {

    outliers[, j] = outliers[, j] && tukey.outlier(x[, j])

  }

  outlier.vec = vector("logical", length = nrow(x))

  for (i in 1:nrow(x)) {

    outlier.vec[i] = all(outliers[i, ])

  }

  return(outlier.vec)

}


Error in outliers[, j] && tukey.outlier(x[, j]) : 

  'length = 10' in coercion to 'logical(1)'

 

What Went Wrong

The problem was on line 4 where I used && instead of &. The double ampersand && only works with single TRUE or FALSE values, but here I was comparing entire vectors of 10 elements. R couldn't handle that and threw an error.

The Fix

I changed && to & for element-wise comparison:

corrected_tukey = function(x) {

  outliers = array(TRUE, dim = dim(x))

  for (j in 1:ncol(x)) {

    outliers[, j] = outliers[, j] & tukey.outlier(x[, j])

  }

  outlier.vec = vector("logical", length = nrow(x))

  for (i in 1:nrow(x)) {

    outlier.vec[i] = all(outliers[i, ])

  }

  return(outlier.vec)

}

Testing the Fix

After fixing the bug, I tested it again:

set.seed(123)

 test_mat = matrix(rnorm(50), nrow = 10) 

corrected_tukey(test_mat)

corrected_tukey(test_mat)

 [1] FALSE FALSE FALSE FALSE FALSE FALSE FALSE

 [8] FALSE FALSE FALSE


The function now works without errors. The output shows that none of the 10 rows have outliers in every column, which makes sense for this random data.

What I Learned

This debugging exercise taught me the difference between logical operators in R. Use && and || for single values, and &and | for vectors. Reading error messages carefully really helps pinpoint where things go wrong.

Comments

Post a Comment

Popular posts from this blog

R Programming Journal - Omar Hamad

Image
  RStudio Installation Experience Yesterday, I switched from Windows to Mac, which was new for me. Installing R and RStudio on macOS  It was simple, but I had to change some security settings in my Mac to allow the software to run.  I downloaded the correct files from the official sites, installed them, and opened RStudio successfully.  What is an R Vector? An R vector is a basic data structure that stores a sequence of elements all of the same type, such as  numbers, words, or logical values. It’s a core part of R because it allows you to organize and work with  data efficiently. Vectors are the foundation for many operations in R, making data manipulation and  analysis faster and easier. Vectors are essential in R because they enable you to perform operations on entire collections of data at  once, rather than handling each element individually. This makes your code more concise and efficient.  Understanding vectors is crucial because most ...
Read more

Assignment 6

Code: # Task 1: Matrix Addition & Subtraction A = matrix(c(2, 0, 1, 3), ncol = 2) B = matrix(c(5, 2, 4, -1), ncol = 2) cat("Matrix A:\n") print(A) cat("\nMatrix B:\n") print(B) A_plus_B = A + B cat("\nA + B:\n") print(A_plus_B) A_minus_B = A - B cat("\nA - B:\n") print(A_minus_B) # Task 2: Create a Diagonal Matrix D = diag(c(4, 1, 2, 3)) cat("\nDiagonal Matrix D:\n") print(D) # Task 3: Construct a Custom 5x5 Matrix first_col = c(3, 2, 2, 2, 2) remaining_block = rbind(   c(1, 1, 1, 1),   diag(3, 4) ) custom_matrix = cbind(first_col, remaining_block) cat("\nCustom 5x5 Matrix:\n") print(custom_matrix) Output: > # Task 1: Matrix Addition & Subtraction > A = matrix(c(2, 0, 1, 3), ncol = 2) > B = matrix(c(5, 2, 4, -1), ncol = 2) > cat("Matrix A:\n") Matrix A: > print(A)      [,1] [,2] [1,]    2    1 [2,]    0    3 > cat("\nMatrix B:\n") Matrix B: > print(B)      [,1] [,2] [1,]   ...
Read more

Generic Functions in R

What I Did   This assignment helped me learn how R's generic functions work and the difference between S3 and S4 objects. I used the mtcars dataset to see how functions change behavior based on what kind of object they're working with.   Loading the Dataset   First, I loaded mtcars data(" mtcars ") head( mtcars ) str( mtcars )   This dataset has 11 columns with things like miles per gallon, horsepower, Etc . How Generic Functions Work   I tested some basic R functions to see how they adapt to different objects:   summary( mtcars ) plot( mtcars )   lm_model = lm (mpg ~ wt + hp, data = mtcars ) summary( lm_model )   What I noticed: when you run summary( ) on a data frame, you get min/max/mean values. But when you run it on a linear model, you get totally different outputs, like coefficients and p-values. Same function name, different behavior!   Creating S3 Objects   S3 is the simpler object system. Here's how I made one:   s3...
Read more