How to Use tapply in R: Apply Functions by Groups Easily

In R, tapply applies a function to subsets of a vector defined by a factor or factors. It splits the data into groups, applies the function to each group, and returns the results in an array or list.

Syntax

The basic syntax of tapply is:

• X: a vector of data values.
• INDEX: a factor or list of factors defining groups.
• FUN: the function to apply to each group.
• ...: optional arguments passed to FUN.
• simplify: logical, whether to simplify the result to an array.

r

tapply(X, INDEX, FUN, ..., simplify = TRUE)

Example

This example shows how to calculate the mean of values grouped by a factor:

r

values <- c(10, 20, 30, 40, 50, 60)
groups <- factor(c("A", "A", "B", "B", "C", "C"))
result <- tapply(values, groups, mean)
print(result)

Output

   A    B    C 
15.0 35.0 55.0 

Common Pitfalls

Common mistakes include:

• Using a non-factor INDEX which can cause unexpected grouping.
• Forgetting that tapply returns an array or list, not a vector.
• Not handling NA values inside the function, which can cause NA results.

Example of a pitfall and fix:

r

# Wrong: INDEX is numeric, not factor
values <- c(1, 2, 3, 4)
groups <- c(1, 1, 2, 2)  # numeric, not factor
result_wrong <- tapply(values, groups, sum)

# Right: convert to factor
groups_factor <- factor(groups)
result_right <- tapply(values, groups_factor, sum)

print(result_wrong)
print(result_right)

Output

1 2 
3 7 
1 2 
3 7 

Quick Reference

Argument	Description
X	Vector of data to split
INDEX	Factor or list of factors to define groups
FUN	Function to apply to each group
...	Additional arguments to FUN
simplify	Whether to simplify output to array (default TRUE)

Key Takeaways

Use tapply to apply a function to subsets of a vector grouped by factors.

Always ensure the grouping variable (INDEX) is a factor for correct grouping.

tapply returns an array or list, not a simple vector.

Handle NA values inside your function to avoid unexpected NA results.

Use the simplify argument to control the output format.