R Error: replacement has length zero — The NA Assignment Bug
Error in x[[i]] <- value : replacement has length zero means you tried to assign a zero-length value (like NULL, character(0), or integer(0)) into a position that expects a real value. This often happens when a lookup or computation silently returns nothing.
Reproducing the Error
x <- list(a = 1, b = 2, c = 3)
# Assigning NULL to a list element with [[ removes it
x[["a"]] <- NULL
cat("After NULL assignment, 'a' is gone:", paste(names(x), collapse = ", "), "\n")
# But assigning NULL to a vector element is an error:
v <- 1:5
tryCatch(
v[[3]] <- NULL,
error = function(e) cat("Error:", conditionMessage(e), "\n")
)
# The more common trigger: a function returns length-0 result
result <- character(0) # Empty character vector
cat("Length of result:", length(result), "\n")
v <- c("a", "b", "c")
tryCatch(
v[[2]] <- result,
error = function(e) cat("Error:", conditionMessage(e), "\n")
)
Cause 1: Lookup Returns Nothing
# Looking up a value that doesn't exist
lookup_table <- c(cat = "meow", dog = "woof", bird = "tweet")
# This returns a named NA — but the name is NA too
animal <- "fish"
sound <- lookup_table[animal]
cat("Lookup result:", sound, "\n")
cat("Is NA:", is.na(sound), "\n")
# With match(), no-match returns NA
idx <- match("fish", names(lookup_table))
cat("match result:", idx, "\n")
# Safe lookup pattern:
safe_lookup <- function(table, key, default = "unknown") {
result <- table[key]
if (is.na(result)) default else unname(result)
}
cat("Safe lookup:", safe_lookup(lookup_table, "fish"), "\n")
cat("Safe lookup:", safe_lookup(lookup_table, "cat"), "\n")
Cause 2: grep/which Returns integer(0)
data <- c("apple", "banana", "cherry")
# grep returns integer(0) when nothing matches
idx <- grep("mango", data)
cat("grep result:", idx, "\n")
cat("Length:", length(idx), "\n")
# Using this empty result for assignment causes issues
results <- character(3)
if (length(idx) > 0) {
results[idx] <- "found"
} else {
cat("No matches — skipping assignment\n")
}
# Pattern: always check length before using grep/which results
matches <- grep("an", data)
cat("Matches for 'an':", matches, "\n")
if (length(matches) > 0) {
cat("Found:", data[matches], "\n")
}
Cause 3: Conditional Assignment with Empty Result
# A function that sometimes returns nothing
get_value <- function(x) {
if (x > 0) return(x * 2)
# Implicitly returns NULL when x <= 0
}
cat("get_value(5):", get_value(5), "\n")
cat("get_value(-1):", get_value(-1), "\n") # NULL
cat("is.null:", is.null(get_value(-1)), "\n")
# Using this in assignment:
results <- numeric(5)
inputs <- c(3, -1, 5, 0, 2)
for (i in seq_along(inputs)) {
val <- get_value(inputs[i])
# Fix: check for NULL before assigning
results[i] <- if (!is.null(val)) val else NA
}
cat("Results:", results, "\n")
Cause 4: Subsetting an Empty Data Frame
df <- data.frame(name = c("Alice", "Bob"), score = c(88, 92))
# Filter returns 0 rows
filtered <- df[df$score > 100, ]
cat("Filtered rows:", nrow(filtered), "\n")
# Trying to extract from empty result
tryCatch(
{
best_name <- filtered$name[[1]]
cat("Best:", best_name, "\n")
},
error = function(e) cat("Error:", conditionMessage(e), "\n")
)
# Fix: check nrow() before extracting
if (nrow(filtered) > 0) {
cat("Best:", filtered$name[[1]], "\n")
} else {
cat("No results matching criteria\n")
}
Prevention Patterns
# Pattern 1: Use %||% (null coalesce) — available in rlang, or define your own
`%||%` <- function(a, b) if (!is.null(a) && length(a) > 0) a else b
result <- NULL %||% "default"
cat("NULL coalesce:", result, "\n")
result2 <- character(0) %||% "fallback"
cat("Empty coalesce:", result2, "\n")
result3 <- "actual_value" %||% "default"
cat("Non-null:", result3, "\n")
# Pattern 2: Defensive function returns
safe_first <- function(x, default = NA) {
if (length(x) == 0 || is.null(x)) return(default)
x[[1]]
}
cat("Normal:", safe_first(c(10, 20)), "\n")
cat("Empty:", safe_first(integer(0)), "\n")
cat("NULL:", safe_first(NULL), "\n")
# Pattern 3: Validate before loop assignment
populate <- function(values, transform) {
results <- vector("list", length(values))
for (i in seq_along(values)) {
out <- transform(values[[i]])
results[[i]] <- if (length(out) > 0) out else NA
}
results
}
filled <- populate(list(5, -1, 3), function(x) if (x > 0) x^2 else NULL)
cat("Populated:", paste(filled, collapse = ", "), "\n")
Practice Exercise
# Exercise: This code crashes when processing certain records.
# Fix it so it handles missing/empty values gracefully.
records <- list(
list(name = "Alice", scores = c(80, 90, 85)),
list(name = "Bob", scores = c()), # Empty scores
list(name = "Carol", scores = c(70, 88)),
list(name = "Dave", scores = NULL) # NULL scores
)
# This crashes:
# for (r in records) {
# avg <- mean(r$scores)
# cat(r$name, ":", avg, "\n")
# }
# Fix it below:
Click to reveal solution
```rrecords <- list(
list(name = "Alice", scores = c(80, 90, 85)),
list(name = "Bob", scores = c()),
list(name = "Carol", scores = c(70, 88)),
list(name = "Dave", scores = NULL)
)
for (r in records) {
if (is.null(r$scores) || length(r$scores) == 0) {
cat(r$name, ": No scores available\n")
} else {
avg <- mean(r$scores)
cat(r$name, ":", round(avg, 1), "\n")
}
}
Summary
| Cause | What Returns Empty | Fix | ||
|---|---|---|---|---|
| Lookup miss | table["nonexistent"] |
Use default: `table[x] % | % default` | |
grep/which no match |
integer(0) |
Check length() > 0 before use |
||
| Function returns NULL | if (cond) value (no else) |
Always include else clause |
||
| Empty filter | df[FALSE, ] |
Check nrow() > 0 |
||
| NULL assignment to vector | v[[i]] <- NULL |
Use NA instead of NULL |
FAQ
What's the difference between NULL, NA, and character(0)?
NULL means "nothing" — the object doesn't exist. NA means "missing value" — a placeholder in a vector. character(0) (or integer(0), numeric(0)) means "empty vector" — it exists but has zero elements. All three can cause "replacement has length zero" when used as assignments.
Why does R allow NULL assignment for lists but not vectors?
In lists, x[["name"]] <- NULL is a valid operation that removes the element. In atomic vectors, every position must have a value (possibly NA), so NULL assignment is meaningless and throws an error.
What's Next?
- R Error: object 'x' not found — the most common R error
- R Error: subscript out of bounds — indexing beyond length
- R Common Errors — the full reference of 50 common errors