How to Use map_dbl(), map_chr(), map_int(), and map_lgl() in R

purrr

purrr map_dbl()

Learn the typed map variants in purrr that return vectors instead of lists. Simplify your functional programming with map_dbl, map_chr, map_int, and map_lgl.

Published

April 3, 2026

Introduction

While map() always returns a list, the typed variants return atomic vectors of a specific type. This is cleaner when you know your function will return a single value of a known type for each element.

The typed map functions: - map_dbl() - returns a numeric (double) vector - map_chr() - returns a character vector - map_int() - returns an integer vector - map_lgl() - returns a logical vector

Getting Started

library(tidyverse)
library(palmerpenguins)

map_dbl(): Return Numeric Values

Use map_dbl() when your function returns a single number:

# Create a list of numeric vectors
numbers <- list(
  a = c(1, 2, 3, 4, 5),
  b = c(10, 20, 30),
  c = c(100, 200, 300, 400)
)

# Calculate mean of each - returns a named numeric vector
map_dbl(numbers, mean)

Practical example with data

# Get mean of each numeric column in penguins
penguins |>
  select(where(is.numeric), -year) |>
  map_dbl(\(x) mean(x, na.rm = TRUE))

Extract model statistics

# Fit models by species and extract R-squared
penguins |>
  drop_na() |>
  split(~species) |>
  map(\(df) lm(body_mass_g ~ flipper_length_mm, data = df)) |>
  map_dbl(\(m) summary(m)$r.squared)

map_chr(): Return Character Values

Use map_chr() when your function returns a single string:

# Get class of each column
penguins |>
  map_chr(class)

Extract names or labels

# Create named list
models <- list(
  simple = lm(mpg ~ wt, data = mtcars),
  multiple = lm(mpg ~ wt + hp, data = mtcars)
)

# Get formula as string
map_chr(models, \(m) deparse(formula(m)))

Format values as strings

# Format numbers with units
measurements <- list(height = 180, weight = 75, age = 30)
units <- c(height = "cm", weight = "kg", age = "years")

map_chr(names(measurements), \(name) {
  paste(measurements[[name]], units[name])
})

map_int(): Return Integer Values

Use map_int() when your function returns a single integer:

# Count elements in each list item
numbers <- list(
  a = 1:5,
  b = 1:10,
  c = 1:3
)

map_int(numbers, length)

Count observations by group

# Count rows in each nested data frame
penguins |>
  drop_na() |>
  split(~species) |>
  map_int(nrow)

Count unique values

# Count unique values in each column
penguins |>
  map_int(\(x) n_distinct(x, na.rm = TRUE))

map_lgl(): Return Logical Values

Use map_lgl() when your function returns TRUE or FALSE:

# Check which columns have missing values
penguins |>
  map_lgl(\(x) any(is.na(x)))

Filter columns programmatically

# Find numeric columns
penguins |>
  map_lgl(is.numeric)

# Use with select
penguins |>
  select(where(\(x) map_lgl(list(x), is.numeric)[[1]]))

Check conditions across list elements

# Check if all values are positive
numbers <- list(a = c(1, 2, 3), b = c(-1, 2, 3), c = c(5, 6, 7))

map_lgl(numbers, \(x) all(x > 0))

Choosing the Right Variant

Use this decision tree:

Does your function return a single value per element?
│
├── No → use map() (returns list)
│
└── Yes → What type?
    ├── Number (double) → map_dbl()
    ├── Integer → map_int()
    ├── Text → map_chr()
    ├── TRUE/FALSE → map_lgl()
    └── Data frame → map_dfr() or list_rbind()

list_rbind() and list_cbind() (purrr 1.0+)

Modern purrr (1.0+) recommends list_rbind() over map_dfr():

# Old way (still works)
map_dfr(1:3, \(i) tibble(x = i, y = i^2))

# New way (purrr 1.0+) - more explicit
map(1:3, \(i) tibble(x = i, y = i^2)) |> list_rbind()

# With ID column
map(1:3, \(i) tibble(x = i, y = i^2)) |>
  list_rbind(names_to = "id")

The new approach is clearer about what’s happening: first map, then combine.

# Column binding
map(1:3, \(i) tibble(!!paste0("col", i) := rnorm(5))) |>
  list_cbind()

Combining map_* with Data Frames

Row-binding results with map_dfr()

When each iteration returns a data frame, use map_dfr() to combine them:

# Get summary stats for each species
penguins |>
  drop_na() |>
  split(~species) |>
  map_dfr(\(df) {
    tibble(
      mean_mass = mean(df$body_mass_g),
      sd_mass = sd(df$body_mass_g),
      n = nrow(df)
    )
  }, .id = "species")

Column-binding with map_dfc()

# Less common, but binds results as columns
map_dfc(1:3, \(i) tibble(!!paste0("col", i) := rnorm(5)))

Common Mistakes

1. Type mismatch errors

# This will error - can't return character from map_dbl
# map_dbl(penguins, class)

# Use map_chr instead
map_chr(penguins, class)

2. Functions returning multiple values

Typed map functions expect exactly one value per element:

# This errors - range() returns 2 values
# map_dbl(numbers, range)

# Solution: return a single value or use map()
map_dbl(numbers, \(x) diff(range(x)))  # Returns range width
map(numbers, range)  # Returns list of ranges

3. NULL values causing errors

# If a function can return NULL, it won't work with typed maps
data_with_nulls <- list(a = 1:3, b = NULL, c = 4:6)

# This errors
# map_int(data_with_nulls, length)

# Handle NULLs first
map_int(data_with_nulls, \(x) if(is.null(x)) 0L else length(x))

Summary

Function	Returns	Use When
`map()`	list	Multiple values or unknown type
`map_dbl()`	numeric vector	Single number per element
`map_chr()`	character vector	Single string per element
`map_int()`	integer vector	Single integer per element
`map_lgl()`	logical vector	TRUE/FALSE per element
`map_dfr()`	data frame	Data frame per element (row-bind)

Typed variants are cleaner than map() |> unlist()
Ensure your function returns exactly one value of the correct type
Handle NULL and NA values explicitly

--- title: "How to Use map_dbl(), map_chr(), map_int(), and map_lgl() in R" description: "Learn the typed map variants in purrr that return vectors instead of lists. Simplify your functional programming with map_dbl, map_chr, map_int, and map_lgl." date: 2026-04-03 categories: ['purrr', 'purrr map_dbl()'] format: html: code-fold: false code-tools: true --- ## Introduction While `map()` always returns a list, the typed variants return atomic vectors of a specific type. This is cleaner when you know your function will return a single value of a known type for each element. **The typed map functions:** - `map_dbl()` - returns a numeric (double) vector - `map_chr()` - returns a character vector - `map_int()` - returns an integer vector - `map_lgl()` - returns a logical vector ## Getting Started ```r library(tidyverse) library(palmerpenguins) ``` ## map_dbl(): Return Numeric Values Use `map_dbl()` when your function returns a single number: ```r # Create a list of numeric vectors numbers <- list( a = c(1, 2, 3, 4, 5), b = c(10, 20, 30), c = c(100, 200, 300, 400) ) # Calculate mean of each - returns a named numeric vector map_dbl(numbers, mean) ``` ### Practical example with data ```r # Get mean of each numeric column in penguins penguins |> select(where(is.numeric), -year) |> map_dbl(\(x) mean(x, na.rm = TRUE)) ``` ### Extract model statistics ```r # Fit models by species and extract R-squared penguins |> drop_na() |> split(~species) |> map(\(df) lm(body_mass_g ~ flipper_length_mm, data = df)) |> map_dbl(\(m) summary(m)$r.squared) ``` ## map_chr(): Return Character Values Use `map_chr()` when your function returns a single string: ```r # Get class of each column penguins |> map_chr(class) ``` ### Extract names or labels ```r # Create named list models <- list( simple = lm(mpg ~ wt, data = mtcars), multiple = lm(mpg ~ wt + hp, data = mtcars) ) # Get formula as string map_chr(models, \(m) deparse(formula(m))) ``` ### Format values as strings ```r # Format numbers with units measurements <- list(height = 180, weight = 75, age = 30) units <- c(height = "cm", weight = "kg", age = "years") map_chr(names(measurements), \(name) { paste(measurements[[name]], units[name]) }) ``` ## map_int(): Return Integer Values Use `map_int()` when your function returns a single integer: ```r # Count elements in each list item numbers <- list( a = 1:5, b = 1:10, c = 1:3 ) map_int(numbers, length) ``` ### Count observations by group ```r # Count rows in each nested data frame penguins |> drop_na() |> split(~species) |> map_int(nrow) ``` ### Count unique values ```r # Count unique values in each column penguins |> map_int(\(x) n_distinct(x, na.rm = TRUE)) ``` ## map_lgl(): Return Logical Values Use `map_lgl()` when your function returns TRUE or FALSE: ```r # Check which columns have missing values penguins |> map_lgl(\(x) any(is.na(x))) ``` ### Filter columns programmatically ```r # Find numeric columns penguins |> map_lgl(is.numeric) # Use with select penguins |> select(where(\(x) map_lgl(list(x), is.numeric)[[1]])) ``` ### Check conditions across list elements ```r # Check if all values are positive numbers <- list(a = c(1, 2, 3), b = c(-1, 2, 3), c = c(5, 6, 7)) map_lgl(numbers, \(x) all(x > 0)) ``` ## Choosing the Right Variant Use this decision tree: ``` Does your function return a single value per element? │ ├── No → use map() (returns list) │ └── Yes → What type? ├── Number (double) → map_dbl() ├── Integer → map_int() ├── Text → map_chr() ├── TRUE/FALSE → map_lgl() └── Data frame → map_dfr() or list_rbind() ``` ## list_rbind() and list_cbind() (purrr 1.0+) Modern purrr (1.0+) recommends `list_rbind()` over `map_dfr()`: ```r # Old way (still works) map_dfr(1:3, \(i) tibble(x = i, y = i^2)) # New way (purrr 1.0+) - more explicit map(1:3, \(i) tibble(x = i, y = i^2)) |> list_rbind() # With ID column map(1:3, \(i) tibble(x = i, y = i^2)) |> list_rbind(names_to = "id") ``` The new approach is clearer about what's happening: first map, then combine. ```r # Column binding map(1:3, \(i) tibble(!!paste0("col", i) := rnorm(5))) |> list_cbind() ``` ## Combining map_* with Data Frames ### Row-binding results with map_dfr() When each iteration returns a data frame, use `map_dfr()` to combine them: ```r # Get summary stats for each species penguins |> drop_na() |> split(~species) |> map_dfr(\(df) { tibble( mean_mass = mean(df$body_mass_g), sd_mass = sd(df$body_mass_g), n = nrow(df) ) }, .id = "species") ``` ### Column-binding with map_dfc() ```r # Less common, but binds results as columns map_dfc(1:3, \(i) tibble(!!paste0("col", i) := rnorm(5))) ``` ## Common Mistakes **1. Type mismatch errors** ```r # This will error - can't return character from map_dbl # map_dbl(penguins, class) # Use map_chr instead map_chr(penguins, class) ``` **2. Functions returning multiple values** Typed map functions expect exactly one value per element: ```r # This errors - range() returns 2 values # map_dbl(numbers, range) # Solution: return a single value or use map() map_dbl(numbers, \(x) diff(range(x))) # Returns range width map(numbers, range) # Returns list of ranges ``` **3. NULL values causing errors** ```r # If a function can return NULL, it won't work with typed maps data_with_nulls <- list(a = 1:3, b = NULL, c = 4:6) # This errors # map_int(data_with_nulls, length) # Handle NULLs first map_int(data_with_nulls, \(x) if(is.null(x)) 0L else length(x)) ``` ## Summary | Function | Returns | Use When | |----------|---------|----------| | `map()` | list | Multiple values or unknown type | | `map_dbl()` | numeric vector | Single number per element | | `map_chr()` | character vector | Single string per element | | `map_int()` | integer vector | Single integer per element | | `map_lgl()` | logical vector | TRUE/FALSE per element | | `map_dfr()` | data frame | Data frame per element (row-bind) | - Typed variants are cleaner than `map() |> unlist()` - Ensure your function returns exactly one value of the correct type - Handle NULL and NA values explicitly ## Related Posts - [How to Use map() in R](/purrr/how-to-use-map-in-r) - [How to Use map2() and pmap() in R](/purrr/how-to-use-map2-and-pmap-in-r) - [How to Use reduce() in R](/purrr/how-to-use-reduce-in-r) - [How to Use across() in R](/dplyr/how-to-use-across-in-r) - [How to Use where() in R](/dplyr/how-to-use-where-in-r)