How to Use separate() in tidyr to Split Columns in R

Use separate() from the tidyr package to split one column into multiple columns by specifying the column to split and the separator. You provide the column name, new column names, and optionally the separator character or regex to split the data cleanly.

Syntax

The basic syntax of separate() is:

• data: Your data frame or tibble.
• col: The name of the column to split.
• into: A character vector of new column names.
• sep: The separator to split on (default is non-alphanumeric).
• remove: Whether to remove the original column (default TRUE).
• convert: Whether to convert the new columns to appropriate types (default FALSE).

r

separate(data, col, into, sep = "[^[:alnum:]]+", remove = TRUE, convert = FALSE)

Example

This example shows how to split a column with names separated by a space into two columns: first and last names.

r

library(tidyr)
library(dplyr)

# Sample data frame
people <- tibble(name = c("John Doe", "Jane Smith", "Alice Johnson"))

# Use separate() to split 'name' into 'first' and 'last'
people_separated <- people %>%
  separate(name, into = c("first", "last"), sep = " ")

print(people_separated)

Output

# A tibble: 3 × 2
  first last   
  <chr> <chr>  
1 John  Doe    
2 Jane  Smith  
3 Alice Johnson

Common Pitfalls

Common mistakes when using separate() include:

• Not specifying enough new column names in into for the splits, causing errors.
• Using the wrong separator sep, so the column does not split as expected.
• Forgetting that separate() removes the original column by default, which might be unwanted.
• Not setting convert = TRUE when you want the new columns to be converted to numbers or factors automatically.

Example of a wrong and right way:

r

library(tidyr)
library(dplyr)

# Wrong: not enough names in 'into'
people <- tibble(name = c("John Doe", "Jane Smith"))

# This will cause an error because 'into' has only one name but two parts exist
# people %>% separate(name, into = c("first"), sep = " ")

# Right: provide two names
people %>% separate(name, into = c("first", "last"), sep = " ")

Output

# A tibble: 2 × 2
  first last 
  <chr> <chr>
1 John  Doe  
2 Jane  Smith

Quick Reference

Argument	Description
data	Data frame or tibble to modify
col	Column name to split
into	Vector of new column names
sep	Separator character or regex (default splits on non-alphanumeric)
remove	Remove original column? (default TRUE)
convert	Convert new columns to appropriate types? (default FALSE)

Key Takeaways

Use separate() to split one column into multiple columns by specifying the column and new names.

Always provide enough new column names in 'into' to match the number of splits.

Specify the correct separator with 'sep' to split data as expected.

Set 'remove = FALSE' if you want to keep the original column after splitting.

Use 'convert = TRUE' to automatically change new columns to proper data types.