Convert Data Frame Column to Numeric in R

In R, converting DataFrame columns to numeric is a common and important step in data preprocessing, particularly for statistical modeling and analysis. Below are effective methods to perform this conversion using base R and the readr package.

Method 1: Convert One Column to Numeric

Start by checking the column to ensure it is in a format that can be converted. If it is factor or character, change it to character with as.character(), and then to numeric with as.numeric(). Employ sapply() to check column classes.

df <- data.frame(
  x1 = as.factor(c(1, 5, 8, 2)),
  x2 = c(3, 2, 5, 2),
  x3 = c(2, 7, 1, 2)
)

# Check current classes
sapply(df, class)

# Convert x1 to numeric
df$x1 <- as.numeric(as.character(df$x1))
sapply(df, class)

Output:

x1 x2 x3
1 1 3 2
2 5 2 7
3 8 5 1
4 2 2 2
x1 x2 x3
"factor" "numeric" "numeric"

Method 2: Convert Multiple Columns to Numeric

To convert multiple columns to numeric, give their indices, then apply() to change each to character and then numeric. Apply sapply() to verify the changes.

# Column indices to be converted
i <- c(2, 3)  
df[, i] <- apply(df[, i], 2, function(x) as.numeric(as.character(x)))
                 
sapply(df, class)

Output:

x1 x2 x3
"numeric" "numeric" "numeric"

Method 3: Using parse_number() from readr

The parse_number() function from the readr package is a useful function to pull numeric values out of character columns. It makes it easy to convert by parsing numbers from strings. The following example uses x1, followed by checking the result.

library(readr)

df <- data.frame(x1 = c("1", "5", "8", "2"),  # Character column
                 x2 = c(3, 2, 5, 2),
                 x3 = c(2, 7, 1, 2))

# Display original data frame
print("Original DataFrame:")
print(df)
print(sapply(df, class))

# Convert using parse_number
df$x1 <- parse_number(df$x1)

# Display modified data frame
print("DataFrame after conversion:")
print(df)

# Verify the class of each column
print("Column Classes:")
print(sapply(df, class))

Output:

[1] "Original DataFrame:"
x1 x2 x3
1 1 3 2
2 5 2 7
3 8 5 1
4 2 2 2
x1 x2 x3
"character" "numeric" "numeric"
[1] "DataFrame after conversion:"
x1 x2 x3
1 1 3 2
2 5 2 7
3 8 5 1
4 2 2 2
[1] "Column Classes:"
x1 x2 x3
"numeric" "numeric" "numeric"