Skip to main content
Talentd

Find the Duplicate Number - Solution & Explanation

MediumArrayTwo PointersBinary SearchBit Manipulation17 min readAsked at: Amazon, Microsoft, Goldman Sachs +19
Practice this problem

Problem Statement

Given an array of integers nums containing n + 1 integers where each integer is in the range [1, n] inclusive.

There is only one repeated number in nums, return this repeated number.

You must solve the problem without modifying the array nums and using only constant extra space.

 

Example 1:

Input: nums = [1,3,4,2,2]
Output: 2

Example 2:

Input: nums = [3,1,3,4,2]
Output: 3

Example 3:

Input: nums = [3,3,3,3,3]
Output: 3

 

Constraints:

  • 1 <= n <= 105
  • nums.length == n + 1
  • 1 <= nums[i] <= n
  • All the integers in nums appear only once except for precisely one integer which appears two or more times.

 

Follow up:

  • How can we prove that at least one duplicate number must exist in nums?
  • Can you solve the problem in linear runtime complexity?

Approach Overview

Problem Overview: You receive an array of n + 1 integers where each value is in the range [1, n]. Since there are more numbers than the range allows, at least one value must appear more than once. The task is to return the duplicate number without modifying the array and while using minimal extra space.

Approach 1: Floyd's Tortoise and Hare (Cycle Detection) (O(n) time, O(1) space)

This approach treats the array like a linked list. Each index points to the next index using the value stored at that position. Because values are restricted to 1..n, following these "next" pointers must eventually form a cycle. The duplicate number is exactly the entry point of that cycle. Use two pointers: a slow pointer moving one step at a time and a fast pointer moving two steps. When they meet, a cycle exists. Reset one pointer to the start and move both one step at a time until they meet again; that position corresponds to the duplicate value. This technique uses the classic two pointers pattern and works without modifying the array. Time complexity is O(n) because each pointer traverses the structure at most a few passes. Space complexity is O(1) since only a couple of variables are used.

Approach 2: Binary Search on Result (O(n log n) time, O(1) space)

This method applies binary search on the value range instead of array indices. The key insight comes from the pigeonhole principle. For any midpoint mid, count how many numbers in the array are less than or equal to mid. If the count exceeds mid, the duplicate must lie in the range [1, mid]. Otherwise, it lies in [mid + 1, n]. Repeatedly narrow the search range until a single value remains. Each step scans the array once to compute the count, giving O(n) work per iteration and O(log n) iterations overall. The algorithm uses constant extra memory and does not modify the input.

Recommended for interviews: Floyd's cycle detection is the solution most interviewers expect. It demonstrates deeper algorithmic thinking because you reinterpret the array as a linked structure and apply cycle detection to locate the repeated value. The binary search approach is also valid and easier to reason about, but it is slower at O(n log n). Showing the binary search idea first and then optimizing to Floyd's algorithm signals strong problem-solving skills.

Approach 1: Approach 1: Floyd's Tortoise and Hare (Cycle Detection)

This approach uses cycle detection to find the duplicate number using two pointers. Imagine the numbers in the array form a linked list, where the value at each index points to the next index. When there is a duplicate, a cycle will be formed. The first step is to find the intersection point of the cycle using a slow and fast pointer. Once they meet, use two pointers to find the entrance of the cycle, which will be the duplicate number. This method doesn't modify the array and uses O(1) space.

This C solution implements the cycle detection algorithm. Two pointers, slow and fast, iterate through the array. First they meet in the cycle; then, by resetting one and moving both pointers step by step, they meet at the duplicate number.

Code

CC++JavaPythonC#JavaScript

C

C++

Java

Python

C#

JavaScript

Complexity

Time Complexity: O(n), where n is the size of the array. Space Complexity: O(1), since we use a constant amount of extra space.

Try this approach in the editor →

Approach 2: Approach 2: Binary Search on Result

This approach leverages a binary search technique on the range of possible values of the duplicate number. For each midpoint in this range, count how many numbers are less than or equal to it. If the count is greater than the midpoint, the duplicate must be in the lower range; otherwise, it is in the upper range. Binary search is applied until the duplicate is found. This doesn't require additional space and doesn't alter the input array.

This C solution uses binary search on the range of possible numbers, counting numbers less than or equal to the midpoint, and then adjusting the binary search boundaries accordingly, eventually discovering the duplicate number.

Code

CC++JavaPythonC#JavaScript

C

C++

Java

Python

C#

JavaScript

Complexity

Time Complexity: O(n log n). Space Complexity: O(1).

Try this approach in the editor →

Approach 3: Binary Search

We can observe that if the number of elements in [1,..x] is greater than x, then the duplicate number must be in [1,..x], otherwise the duplicate number must be in [x+1,..n].

Therefore, we can use binary search to find x, and check whether the number of elements in [1,..x] is greater than x at each iteration. This way, we can determine which interval the duplicate number is in, and narrow down the search range until we find the duplicate number.

The time complexity is O(n times log n), where n is the length of the array nums. The space complexity is O(1).

Code

PythonJavaC++GoTypeScriptRustJavaScript

Python

Java

C++

Go

TypeScript

Rust

JavaScript

Try this approach in the editor →

Complexity Comparison

ApproachComplexity
Approach 1: Floyd's Tortoise and Hare (Cycle Detection)

Time Complexity: O(n), where n is the size of the array. Space Complexity: O(1), since we use a constant amount of extra space.

Approach 2: Binary Search on Result

Time Complexity: O(n log n). Space Complexity: O(1).

Binary Search—

Detailed Complexity Analysis

ApproachTimeSpaceWhen to Use
Floyd's Tortoise and Hare (Cycle Detection)O(n)O(1)Optimal interview solution when array cannot be modified and constant space is required
Binary Search on Value RangeO(n log n)O(1)Useful when you recognize the pigeonhole principle and want a simpler reasoning approach

Video Solution

Find the Duplicate Number - Floyd's Cycle Detection - Leetcode 287 - Python • NeetCode • 464,747 views views

Watch 9 more video solutions →

Frequently Asked Questions

Is Find the Duplicate Number easy or hard?
Find the Duplicate Number is rated Medium difficulty. The brute reasoning is straightforward, but recognizing that the array forms a cycle and applying Floyd's algorithm requires deeper algorithmic insight.
Find the Duplicate Number Python/Java solution
Python and Java implementations usually follow Floyd's cycle detection pattern: run a slow and fast pointer until they meet, then reset one pointer to the start and move both step by step to locate the duplicate. The algorithm uses constant space and runs in linear time.
How to solve Find the Duplicate Number in O(n)?
Use Floyd's Tortoise and Hare algorithm. Move a slow pointer by one step and a fast pointer by two steps through indices defined by array values until they meet. Reset one pointer to the start and move both one step at a time; the meeting point identifies the duplicate number.
What is the best approach for Find the Duplicate Number?
Floyd's Tortoise and Hare cycle detection algorithm is the optimal approach. It runs in O(n) time and O(1) space while keeping the array unchanged. The trick is modeling the array as a linked list where each value points to the next index, which naturally forms a cycle at the duplicate.
Is Find the Duplicate Number asked at Google/Amazon/Meta?
Find the Duplicate Number frequently appears in interviews at companies like Google, Amazon, and Meta. It tests understanding of cycle detection, pointer techniques, and reasoning about constraints such as not modifying the input array and using constant extra space.
What data structure is used in Find the Duplicate Number?
The problem uses an array as the input structure, but the optimal solution conceptually treats it like a linked list. Each index points to another index based on its value, allowing cycle detection techniques with two pointers to identify the duplicate.
What is the time complexity of Find the Duplicate Number?
The optimal solution using Floyd's cycle detection runs in O(n) time with O(1) extra space. Another common method uses binary search on the value range, which takes O(n log n) time because each binary search step requires scanning the entire array.

Ready to solve this problem?

Practice Find the Duplicate Number with our built-in code editor and test cases.

Practice on FleetCode
Previous

First Missing Positive

Next

Linked List Cycle II