What it means. R looked for an object named x in the current environment and could not find it. The object either was never created, was created in a different session, or its name is misspelled.

Most common causes.

# Cause 1: typo in object name
temperature_data <- c(12, 14, 11)
mean(temperture_data)   # Error: object 'temperture_data' not found

# Cause 2: chunk ran out of order in RMarkdown — object from chunk 5
# is referenced in chunk 3, which runs first during knit
# Fix: re-order your chunks so dependencies flow top to bottom

# Cause 3: object created inside an if() or for() block with wrong scope
for (i in 1:5) { result_i <- i^2 }
result_i   # Works — last iteration's value is in global env
# But if the for loop never ran (e.g., empty vector), result_i won't exist

# Cause 4: package function used before library() call
mutate(df, x = x * 2)   # Error if dplyr is not loaded

Fix. Run ls() to list objects in your current environment. Check spelling. In RMarkdown, ensure the chunk that creates the object runs before the chunk that uses it.

What it means. R could not find a function named x. Unlike “object not found” (which applies to data), this applies to functions — it almost always means a package is not loaded.

Most common causes.

# Cause 1: forgot library()
ggplot(df, aes(x = site, y = do_mgl)) + geom_col()
# Error: could not find function "ggplot"
# Fix:
library(ggplot2)   # or library(tidyverse)

# Cause 2: typo in function name
raed_csv("data/wq.csv")       # should be read_csv
filter(df, siet == "Upland")  # filter is fine, but check the column name typo too

# Cause 3: package installed but namespace not loaded
vegan::rarefy(comm, sample = 50)   # This works without library(vegan)
rarefy(comm, sample = 50)           # Error — function not in search path

Fix. Add library(packagename) near the top of your script, before any code that uses the package. Check spelling with ?function_name — if the help page opens, the function exists.

What it means. R tried to read a file and could not find it at the path you specified.

Most common causes.

# Cause 1: wrong working directory
# R is looking in the wrong folder
read.csv("data/water_quality.csv")
# Fix: check and set your working directory
getwd()
setwd("/correct/path/to/project")
# Better: use RStudio Projects so working directory is always the project root

# Cause 2: path is relative to the wrong location
# In RMarkdown, the working directory during knit is the .Rmd file's location,
# not your console working directory
read.csv("../data/water_quality.csv")   # go up one level, then into data/

# Cause 3: typo in filename (including wrong case — R is case-sensitive on Linux/Mac)
read.csv("data/Water_Quality.csv")   # file is actually water_quality.csv

Fix. Run list.files("data/") to see exactly what is in the folder. Use file.exists("data/water_quality.csv") to test a specific path before reading.

# Defensive pattern
path <- "data/water_quality.csv"
if (!file.exists(path)) stop("File not found: ", path)
df <- read.csv(path)

What it means. You tried to use [ or $ on a function rather than a data frame or vector. A “closure” is R’s technical term for a function.

Most common cause. A variable name shadows a function name — you used data, df, filter, mean, or another common name for your data object and for a function, and R found the function first.

# Cause: common variable names that shadow functions
data <- read.csv("myfile.csv")   # 'data' is also a base R function
data$site                        # Error: object of type 'closure' is not subsettable

# Fix: use a more specific variable name
wq_data <- read.csv("myfile.csv")
wq_data$site   # works fine

# Other names to avoid as variable names:
# c, t, T, F, mean, sum, min, max, df, data, matrix, list, table

Fix. Rename your variable to something that does not clash with a built-in function. Run is.function(data) — if it returns TRUE, data is being interpreted as the function.

What it means. You tried to assign a vector to a data frame column, but the vector length does not match the number of rows.

wq <- data.frame(site = c("A","B","C","D"), do_mgl = c(7,8,6,9))

# Cause 1: vector is the wrong length
wq$log_do <- log(c(7, 8))   # Error: replacement has 2 rows, data has 4

# Cause 2: filter() or subset() happened upstream and changed nrow
wq_sub <- wq[wq$do_mgl > 7, ]   # now 2 rows
wq$flag <- wq_sub$do_mgl > 7    # wq still has 4 rows — mismatch

# Fix: always operate on the same object
wq$log_do <- log(wq$do_mgl)           # same-length operation
wq$flag   <- wq$do_mgl > 7            # correct: compares within the same df

Fix. Check length() of the vector and nrow() of the data frame before assigning. When in doubt, use mutate() inside a dplyr pipeline — it always operates on the rows of the piped object.

What it means. You tried to do arithmetic (+, -, *, /, ^) on something that is not a number — most often a character string or a factor.

wq <- data.frame(
  site   = c("A","B","C"),
  temp   = c("12.1", "14.5", "11.8")   # stored as character, not numeric!
)

mean(wq$temp)        # returns NA with a warning
wq$temp_scaled <- wq$temp * 0.1  # Error: non-numeric argument to binary operator

Diagnosis and fix.

class(wq$temp)      # "character" — that's the problem
str(wq)             # shows all column types at once

# Fix 1: convert the column
wq$temp <- as.numeric(wq$temp)

# Fix 2: fix at import (readr does this automatically for properly formatted CSVs)
library(readr)
wq <- read_csv("wq.csv")   # detects column types automatically

# Fix 3: check for stray characters that prevent coercion
# If as.numeric() produces NAs, the column contains non-numeric strings
wq$temp[is.na(as.numeric(wq$temp))]   # shows the problematic values

What it means. You used [ with a column name or number that does not exist in the data frame.

df <- data.frame(site = "A", temp_C = 12, do_mgl = 8)

df[, "temperature"]   # Error: undefined columns selected — column is "temp_C"
df[, 5]               # Error: there is no 5th column

# With $, the behavior is different — silently returns NULL
df$temperature        # NULL (no error!)

Fix. Use names(df) or colnames(df) to check exact column names. The $ operator’s silent NULL return for missing columns is more dangerous than this error, because it propagates invisibly through your pipeline.

# Defensive check
col <- "temperature"
if (!col %in% names(df)) stop("Column '", col, "' not found in data frame")

What it means. You tried to create a data.frame() from vectors of different lengths, and none of them is a multiple of the others (so recycling cannot resolve it).

data.frame(
  site  = c("A", "B", "C"),
  temp  = c(12.1, 14.5),       # only 2 values — Error
  do_mg = c(7.2, 8.1, 6.9)
)

Fix. All vectors passed to data.frame() must have the same length. Check length() on each vector before constructing the frame. The error message tells you the conflicting lengths.

What it means. You used $ on a plain vector (numeric, character, or logical), not a list or data frame. $ only works on named list-like objects.

# Cause 1: trying to use $ on a vector returned by a function
result <- c(mean = 12.4, sd = 2.1)   # named numeric vector, not a list
result$mean   # Error: $ operator is invalid for atomic vectors

# Fix: use [[ ]] or named indexing
result[["mean"]]    # 12.4
result["mean"]      # returns named vector

# Cause 2: function returns a vector, not a data frame
summary_values <- tapply(wq$temp, wq$site, mean)
summary_values$A    # Error: tapply returns an array, not a data frame
summary_values["A"] # Correct

What it means. The package was compiled for a newer (or older) version of R than you are running. This is a warning, not an error — the package usually still loads and works. It becomes a problem if the package uses features that do not exist in your R version.

Fix. Update R and/or the package. Run update.packages() to update all installed packages. If you need a specific older version for reproducibility, use renv.

What it means. You called library(X) but package X is not installed on this machine.

library(vegan)
# Error in library(vegan) : there is no package called 'vegan'

Fix.

install.packages("vegan")   # installs from CRAN
library(vegan)              # then load it

# For GitHub-only packages:
# install.packages("remotes")
# remotes::install_github("username/packagename")

In RMarkdown, include install lines commented out with a note:

# Run this line once to install, then comment it out:
# install.packages("vegan")
library(vegan)

What it means. as.numeric() (or as.integer()) could not convert one or more values — they became NA. This almost always means your column contains non-numeric strings mixed in with numbers.

x <- c("12.1", "14.5", "N/A", "11.8", "--", "10.2")
as.numeric(x)
# [1] 12.1 14.5   NA 11.8   NA 10.2
# Warning: NAs introduced by coercion

This is how you clean sensor data with flags. Find and inspect the non-numeric values, decide whether to recode them, then convert.

# Find which values will not convert
x[is.na(suppressWarnings(as.numeric(x)))]
# [1] "N/A" "--"

# Recode to NA explicitly, then convert
x_clean <- ifelse(x %in% c("N/A", "--", ""), NA_character_, x)
as.numeric(x_clean)   # no warning; NAs are intentional

What it means. You took the log (or sqrt) of a negative number or zero. log(0) returns -Inf; log(-1) returns NaN (Not a Number). Both propagate silently through arithmetic.

conc <- c(0.5, 1.2, 0.0, -0.1, 2.3)   # zero and negative from a sensor glitch
log(conc)
# [1] -0.6931  0.1823    -Inf    NaN  0.8329
# Warning message: In log(conc) : NaNs produced

Fix. Investigate why negatives or zeros exist. Sensor values of 0 or below the detection limit need special handling (see the Non-Detects module). Filter or recode before transforming.

conc_clean <- conc[conc > 0]   # remove zeros and negatives
log(conc_clean)                # no warning

What it means. You performed a vectorized operation on two vectors of different lengths, and R recycled the shorter vector to match the longer one — but the lengths are not evenly divisible. The recycling probably produced wrong results.

a <- c(1, 2, 3, 4, 5, 6, 7)
b <- c(10, 20, 30)
a + b   # Warning: longer object length is not a multiple of shorter
# [1] 11 22 33 14 25 36 17
# R recycled b as 10, 20, 30, 10, 20, 30, 10 — almost certainly not intended

This warning signals a logic error, not a data glitch. Check that your vectors are the lengths you intend.

What it means. ggplot2 encountered NA values in the columns mapped to x or y aesthetics. It silently dropped those rows from the plot. This is often expected behavior — but always ask whether the NAs are intentional or indicate a parsing error upstream.

# Diagnosis: how many NAs are there and where?
sum(is.na(df$do_mgl))
which(is.na(df$do_mgl))   # which rows

# If the NAs came from a failed date parse, for example,
# fix the parse rather than just suppressing the ggplot warning

What it means. lubridate’s ymd(), mdy(), etc. could not parse one or more values — they became NA. This is the single most important warning in time-series analysis because it means you have lost data points silently.

dates <- c("2024-01-15", "2024-02-20", "20240330", "bad date")
ymd(dates)
# [1] "2024-01-15" "2024-02-20" "2024-03-30" NA
# Warning message: 1 failed to parse.

Fix. Find and inspect the failed values before proceeding.

parsed <- ymd(dates, quiet = TRUE)
dates[is.na(parsed)]   # "bad date"
# Fix, recode, or document the bad value, then re-parse