Fix randomForest.default(m, y, …) : Na/NaN/Inf in foreign function call

---

One error you may encounter in R is:

Error in randomForest.default(m, y, ...) : 
  NA/NaN/Inf in foreign function call (arg 1)

There are two reasons for why this error might occur:

• There are NA, NaN, or Inf values in the dataset
• One of the variables in the dataset is a character

The easiest way to fix this error is to remove rows with missing data and convert character variables to factor variables:

#remove rows with missing values 
df <- na.omit(df)

#convert all character variables to factor variables
library(dplyr)
df %>% mutate_if(is.character, as.factor)

This tutorial shares an example of how to fix this error in practice.

Related: 

How to Reproduce the Error

Suppose we attempt to fit a random forest to the following data frame in R:

library(randomForest)

#create data frame
df <- data.frame(y <- c(30, 29, 30, 45, 23, 19, 9, 8, 11, 14),
                 x1 <- c('A', 'A', 'B', 'B', 'B', 'B', 'C', 'C', 'C', 'C'),
                 x2 <- c(4, 4, 5, 7, 8, 7, 9, 6, 13, 15))

#attempt to fit random forest model
model <- randomForest(formula = y ~ ., data = df)

Error in randomForest.default(m, y, ...) :
  NA/NaN/Inf in foreign function call (arg 1)

We receive an error because x1 is a character variable in the data frame.

We can confirm this by using the str() function to view the structure of the data frame:

str(df)

'data.frame':	10 obs. of  3 variables:
 $ y....c.30..29..30..45         : num  30 29 30 45 23 19 9 8 11 14
 $ x1....c..A....A....B....B.... : chr  "A" "A" "B" "B"
 $ x2....c.4..4..5..7..          : num  4 4 5 7 8 7 9 6 13 15

How to Fix the Error

To fix this error, we can use the mutate_if() function from to convert each character column to a factor column:

library(dplyr)

#convert each character column to factor
df = df %>% mutate_if(is.character, as.factor)

#fit random forest model
model <- randomForest(formula = y ~ ., data = df)

#view summary of model
model

Call:
 randomForest(formula = y ~ ., data = df) 
               Type of random forest: regression
                     Number of trees: 500
No. of variables tried at each split: 1

          Mean of squared residuals: 65.0047
                    % Var explained: 48.64

We don’t receive any error this time because there are no longer any character variables in the data frame.

The following tutorials explain how to address other common errors in R: