How to reshape data from wide to long with pivot_longer() in R

tidyr

tidyr pivot_longer()

Learn reshape data from wide to long with pivot_longer() in r with clear examples and explanations.

Published

March 26, 2026

Introduction

The pivot_longer() function from the tidyr package transforms wide data into long format by gathering multiple columns into key-value pairs. This is essential for data analysis and visualization, as most R functions expect data in “tidy” format where each variable forms a column and each observation forms a row. Use pivot_longer() when you have data spread across multiple columns that should be gathered into a single variable.

Basic Setup

First, let’s load the required packages for data manipulation:

library(tidyverse)

Simple Example: Basic Pivot

Let’s start with a simple dataset containing patient measurements at two time points:

df <- tibble(pt_id = 1:4,
             h1 = c(80, 82, 85, 88),
             h2 = c(100, 102, 105, 110))
df

This creates a wide-format dataset where each patient has measurements in separate columns (h1 and h2).

Now we’ll pivot this data from wide to long format using pivot_longer():

df |>
  pivot_longer(cols = h1:h2, 
               names_to = "measurement", 
               values_to = "value")

The result transforms our 4-row dataset into an 8-row dataset where each measurement becomes a separate observation. The cols argument specifies which columns to pivot, names_to creates a new column with the original column names, and values_to stores the actual values.

Advanced Example: Separating Column Names

Often, column names contain multiple pieces of information. Here’s a dataset where column names include both measurement time and gender:

df <- tibble(pt_id = 1:4,
             h1_f = c(80, 82, 85, 88),
             h2_m = c(100, 102, 105, 110))
df

This dataset has column names that combine measurement time (h1, h2) with gender information (f, m).

We can separate these components during the pivot operation:

df |>
  pivot_longer(cols = h1_f:h2_m, 
               names_sep = "_",
               names_to = c("measurement", "sex"),
               values_to = "value")

The names_sep argument tells pivot_longer() to split column names at the underscore, creating two new columns: “measurement” and “sex”. This creates a fully tidy dataset where each piece of information has its own column.

Summary

The pivot_longer() function is crucial for reshaping data from wide to long format. Use the basic form with cols, names_to, and values_to for simple pivoting, and add names_sep when column names contain multiple variables separated by a delimiter. This transformation is often the first step in creating tidy datasets ready for analysis and visualization.

--- title: "How to reshape data from wide to long with pivot_longer() in R" description: "Learn reshape data from wide to long with pivot_longer() in r with clear examples and explanations." date: 2026-03-26 categories: ['tidyr', 'tidyr pivot_longer()'] format: html: code-fold: false code-tools: true --- ## Introduction The `pivot_longer()` function from the tidyr package transforms wide data into long format by gathering multiple columns into key-value pairs. This is essential for data analysis and visualization, as most R functions expect data in "tidy" format where each variable forms a column and each observation forms a row. Use `pivot_longer()` when you have data spread across multiple columns that should be gathered into a single variable. ## Basic Setup First, let's load the required packages for data manipulation: ```r library(tidyverse) ``` ## Simple Example: Basic Pivot Let's start with a simple dataset containing patient measurements at two time points: ```r df <- tibble(pt_id = 1:4, h1 = c(80, 82, 85, 88), h2 = c(100, 102, 105, 110)) df ``` This creates a wide-format dataset where each patient has measurements in separate columns (h1 and h2). Now we'll pivot this data from wide to long format using `pivot_longer()`: ```r df |> pivot_longer(cols = h1:h2, names_to = "measurement", values_to = "value") ``` The result transforms our 4-row dataset into an 8-row dataset where each measurement becomes a separate observation. The `cols` argument specifies which columns to pivot, `names_to` creates a new column with the original column names, and `values_to` stores the actual values. ## Advanced Example: Separating Column Names Often, column names contain multiple pieces of information. Here's a dataset where column names include both measurement time and gender: ```r df <- tibble(pt_id = 1:4, h1_f = c(80, 82, 85, 88), h2_m = c(100, 102, 105, 110)) df ``` This dataset has column names that combine measurement time (h1, h2) with gender information (f, m). We can separate these components during the pivot operation: ```r df |> pivot_longer(cols = h1_f:h2_m, names_sep = "_", names_to = c("measurement", "sex"), values_to = "value") ``` The `names_sep` argument tells `pivot_longer()` to split column names at the underscore, creating two new columns: "measurement" and "sex". This creates a fully tidy dataset where each piece of information has its own column. ## Summary The `pivot_longer()` function is crucial for reshaping data from wide to long format. Use the basic form with `cols`, `names_to`, and `values_to` for simple pivoting, and add `names_sep` when column names contain multiple variables separated by a delimiter. This transformation is often the first step in creating tidy datasets ready for analysis and visualization.