Stopping Scripts in R

I use a series of scripts to generate regular reports, and I often want the process to pause at key points—either to check progress or to flag issues that need my attention. One common example involves a lookup table that maps each fundraiser to their team. I append that team information to my reports so I can easily segment the data by team and individual fundraiser.

But what happens when something changes? A new team member joins and isn’t in the lookup table yet. Someone gets married and their last name changes. Suddenly, that person is no longer associated with the correct team, and the report quietly becomes wrong. Here’s how I catch those issues before they slip through.

Build the Data

Portfolio

# Load the dplyr library
library(dplyr)

# Build a sample portfolio
df <- tibble(fundraiser = c("Carol", 
                            "Bob", 
                            "Rachel"),
             constituent = c("Charles Dickens", 
                             "Victor Hugo", 
                             "Jane Austen")
             )

# Look at the data
df

# A tibble: 3 × 2
  fundraiser constituent    
  <chr>      <chr>          
1 Carol      Charles Dickens
2 Bob        Victor Hugo    
3 Rachel     Jane Austen    

Lookup Table

# Create a fundraiser lookup table
lookup <- tibble(fundraiser = c("Carol", "Rachel"),
                 team = c("Major Giving", "Annual Giving"))

# Look at the data
lookup

# A tibble: 2 × 2
  fundraiser team         
  <chr>      <chr>        
1 Carol      Major Giving 
2 Rachel     Annual Giving

Join the Data

Now, let’s see what happens when I join lookup to df.

df <- left_join(df, lookup, by = c("fundraiser" = "fundraiser"))

df

# A tibble: 3 × 3
  fundraiser constituent     team         
  <chr>      <chr>           <chr>        
1 Carol      Charles Dickens Major Giving 
2 Bob        Victor Hugo     <NA>         
3 Rachel     Jane Austen     Annual Giving

If I wasn’t paying attention, I might keep going, produce a beautifully formatted report, save it to Excel, and only then realize that Bob isn’t connected with a team.

Stopping the Script

There are a few ways to have R stop the script. The basic structure is if a condition exists, then do something. Here, we check for empty strings or NA’s in the data and, if they’re found, stop and display the message.

if(any(df$team == "" | is.na(df$team))) {
  stop(paste("Someone is missing a team. Does the lookup table need to be updated?"))
}

Error:
! Someone is missing a team. Does the lookup table need to be updated?

This is a simple way to throw an error if something doesn’t look right, but we can take it a step further by listing the names of the fundraisers who are missing a team.

if(any(df$team == "" | is.na(df$team))) {
  stop(
    paste(
      "These fundraisers have no team:",
      paste(df$fundraiser[df$team == "" | is.na(df$team)], collapse = ", ")
    )
  )
}

Error:
! These fundraisers have no team: Bob

Wrapping it in a Function

It is conceiveable that we might want to run similar checks on other columns in the data. Here is a function that does the same process, but wrapped in a more flexible package that can be pointed at any data set, name column, and target column.

check_missing <- function(data, name_col, target_col) {
  
  # Convert column names to symbols
  name_col <- rlang::ensym(name_col)
  target_col <- rlang::ensym(target_col)
  
  # Identify rows where the target column is blank or NA
  missing_rows <- data |> 
    dplyr::filter(is.na(!!target_col) | !!target_col == "") |> 
    dplyr::pull(!!name_col)
  
  # If any are missing, stop with a helpful message
  if(length(missing_rows) > 0) {
    stop(
      paste(
        "These entries have missing values in", rlang::as_string(target_col), "-", paste(missing_rows, collapse = ", ")
      )
    )
  }
  invisible(TRUE)
}

check_missing(df, fundraiser, team)

Error in `check_missing()`:
! These entries have missing values in team - Bob