Debugging

2026-03-03

Debugging

Fixing problems

Main Tools

• traceback() - the sequence of calls that led to the error
  • runs after you’ve gotten the error
• Setting breakpoints
• browser() - open an interactive session at an arbitrary breakpoint
  • can be used to create conditional breakpoints
• debug() - open an interactive session where the error occurred
  • RStudio has a rerun with debug option

Or … low tech

• print() statements
• going line-by-line through the code interactively

Useful References: - Chapters Condition handling and Debugging by Hadley Wickham - RStudio Debugging Tools

First go to when you get an unexpected result

Placing output such as

• cat()
• str()
• head()
• summary()
• print()

in strategic places in functions, are simple ways to build code up step-wise ensuring that each piece works (and figuring out where it fails)

Example: Read code

What does this function do? How does it work?

somefunction <- function (...) 
{
    k <- length(ll <- list(...))
    if (k == 0L) 
        return(invisible())
    mc <- match.call()
    for (i in 1L:k) if (!(is.logical(r <- ll[[i]]) && !any(is.na(r)) && 
        all(r))) {
        ch <- deparse(mc[[i + 1]], width.cutoff = 60L)
        if (length(ch) > 1L) 
            ch <- paste(ch[1L], "....")
        stop(paste(ch, " is not ", if (length(r) > 1L) 
            "all ", "TRUE", sep = ""), call. = FALSE)
    }
    invisible()
}

Step-by-step code checking

The browser() function allows stepping through your code.

help(browser)

• n= executes the next line of code

• s= step into next line. Same as n except if the line is a function. Will step into function

• f = finish execution of this loop

• c= continue execution until the end of the function, and exit

• where gives you the stack information, the trace of the active function calls

• Q exits browser()

Let’s try browser()

somefunction2() function has browser() in first line of function

somefunction2 <- function (...) 
{
    browser()
    k <- length(ll <- list(...))
    if (k == 0L) 
        return(invisible())
    mc <- match.call()
    for (i in 1L:k) if (!(is.logical(r <- ll[[i]]) && !any(is.na(r)) && 
        all(r))) {
        ch <- deparse(mc[[i + 1]], width.cutoff = 60L)
        if (length(ch) > 1L) 
            ch <- paste(ch[1L], "....")
        stop(paste(ch, " is not ", if (length(r) > 1L) 
            "all ", "TRUE", sep = ""), call. = FALSE)
    }
    invisible()
}

x <- 1
somefunction2(x==x, 1+1==2, c(7+5==11.99999, 1+1==2))

Browser enables:

• printing of result of each line

• checking changes to values

• make sure that it results in what is expected

debug()

• debug(f)
  automatically places a browser() statement in the first line of function f, i.e. browser starts every time the function f is being executed

• undebug(f)
  removes it.

• debugonce(f)
  Convenient alternative: starts the browser the next time function f is being executed

Setting error handling

• options(error=recover) will drop user into browser() upon an error

• options(error=NULL) default, do nothing

• options(warn=2) sets maximum number of warnings to be 2

• options(warn=0) warnings are stored until top level function finishes (default)

• options(warn=-1) ignore all warnings

Your Turn: fix this!!

larger(x,y) is supposed to return the element-wise maximum of two vectors

larger <- function(x, y) { 
  y.is.bigger <- y > x 
    x[y.is.bigger] <- y[y.is.bigger] 
    x
} 
larger(c(1, 5, 10), c(2, 4, 11)) 
larger(c(1, 5, 10), 6)

why is there an NA in the second example? It should be a 6. Figure out why this happens, and try to fix it.