{ "data": { "question": { "questionId": "480", "questionFrontendId": "480", "boundTopicId": null, "title": "Sliding Window Median", "titleSlug": "sliding-window-median", "content": "

The median is the middle value in an ordered integer list. If the size of the list is even, there is no middle value. So the median is the mean of the two middle values.

\n\n\n\n

You are given an integer array nums and an integer k. There is a sliding window of size k which is moving from the very left of the array to the very right. You can only see the k numbers in the window. Each time the sliding window moves right by one position.

\n\n

Return the median array for each window in the original array. Answers within 10-5 of the actual value will be accepted.

\n\n

 

\n

Example 1:

\n\n
\nInput: nums = [1,3,-1,-3,5,3,6,7], k = 3\nOutput: [1.00000,-1.00000,-1.00000,3.00000,5.00000,6.00000]\nExplanation: \nWindow position                Median\n---------------                -----\n[1  3  -1] -3  5  3  6  7        1\n 1 [3  -1  -3] 5  3  6  7       -1\n 1  3 [-1  -3  5] 3  6  7       -1\n 1  3  -1 [-3  5  3] 6  7        3\n 1  3  -1  -3 [5  3  6] 7        5\n 1  3  -1  -3  5 [3  6  7]       6\n
\n\n

Example 2:

\n\n
\nInput: nums = [1,2,3,4,2,3,1,4,2], k = 3\nOutput: [2.00000,3.00000,3.00000,3.00000,2.00000,3.00000,2.00000]\n
\n\n

 

\n

Constraints:

\n\n\n", "translatedTitle": null, "translatedContent": null, "isPaidOnly": false, "difficulty": "Hard", "likes": 3079, "dislikes": 183, "isLiked": null, "similarQuestions": "[{\"title\": \"Find Median from Data Stream\", \"titleSlug\": \"find-median-from-data-stream\", \"difficulty\": \"Hard\", \"translatedTitle\": null}]", "exampleTestcases": "[1,3,-1,-3,5,3,6,7]\n3\n[1,2,3,4,2,3,1,4,2]\n3", "categoryTitle": "Algorithms", "contributors": [], "topicTags": [ { "name": "Array", "slug": "array", "translatedName": null, "__typename": "TopicTagNode" }, { "name": "Hash Table", "slug": "hash-table", "translatedName": null, "__typename": "TopicTagNode" }, { "name": "Sliding Window", "slug": "sliding-window", "translatedName": null, "__typename": "TopicTagNode" }, { "name": "Heap (Priority Queue)", "slug": "heap-priority-queue", "translatedName": null, "__typename": "TopicTagNode" } ], "companyTagStats": null, "codeSnippets": [ { "lang": "C++", "langSlug": "cpp", "code": "class Solution {\npublic:\n vector medianSlidingWindow(vector& nums, int k) {\n \n }\n};", "__typename": "CodeSnippetNode" }, { "lang": "Java", "langSlug": "java", "code": "class Solution {\n public double[] medianSlidingWindow(int[] nums, int k) {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Python", "langSlug": "python", "code": "class Solution(object):\n def medianSlidingWindow(self, nums, k):\n \"\"\"\n :type nums: List[int]\n :type k: int\n :rtype: List[float]\n \"\"\"\n ", "__typename": "CodeSnippetNode" }, { "lang": "Python3", "langSlug": "python3", "code": "class Solution:\n def medianSlidingWindow(self, nums: List[int], k: int) -> List[float]:\n ", "__typename": "CodeSnippetNode" }, { "lang": "C", "langSlug": "c", "code": "/**\n * Note: The returned array must be malloced, assume caller calls free().\n */\ndouble* medianSlidingWindow(int* nums, int numsSize, int k, int* returnSize) {\n \n}", "__typename": "CodeSnippetNode" }, { "lang": "C#", "langSlug": "csharp", "code": "public class Solution {\n public double[] MedianSlidingWindow(int[] nums, int k) {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "JavaScript", "langSlug": "javascript", "code": "/**\n * @param {number[]} nums\n * @param {number} k\n * @return {number[]}\n */\nvar medianSlidingWindow = function(nums, k) {\n \n};", "__typename": "CodeSnippetNode" }, { "lang": "TypeScript", "langSlug": "typescript", "code": "function medianSlidingWindow(nums: number[], k: number): number[] {\n \n};", "__typename": "CodeSnippetNode" }, { "lang": "PHP", "langSlug": "php", "code": "class Solution {\n\n /**\n * @param Integer[] $nums\n * @param Integer $k\n * @return Float[]\n */\n function medianSlidingWindow($nums, $k) {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Swift", "langSlug": "swift", "code": "class Solution {\n func medianSlidingWindow(_ nums: [Int], _ k: Int) -> [Double] {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Kotlin", "langSlug": "kotlin", "code": "class Solution {\n fun medianSlidingWindow(nums: IntArray, k: Int): DoubleArray {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Dart", "langSlug": "dart", "code": "class Solution {\n List medianSlidingWindow(List nums, int k) {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Go", "langSlug": "golang", "code": "func medianSlidingWindow(nums []int, k int) []float64 {\n \n}", "__typename": "CodeSnippetNode" }, { "lang": "Ruby", "langSlug": "ruby", "code": "# @param {Integer[]} nums\n# @param {Integer} k\n# @return {Float[]}\ndef median_sliding_window(nums, k)\n \nend", "__typename": "CodeSnippetNode" }, { "lang": "Scala", "langSlug": "scala", "code": "object Solution {\n def medianSlidingWindow(nums: Array[Int], k: Int): Array[Double] = {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Rust", "langSlug": "rust", "code": "impl Solution {\n pub fn median_sliding_window(nums: Vec, k: i32) -> Vec {\n \n }\n}", "__typename": "CodeSnippetNode" }, { "lang": "Racket", "langSlug": "racket", "code": "(define/contract (median-sliding-window nums k)\n (-> (listof exact-integer?) exact-integer? (listof flonum?))\n )", "__typename": "CodeSnippetNode" }, { "lang": "Erlang", "langSlug": "erlang", "code": "-spec median_sliding_window(Nums :: [integer()], K :: integer()) -> [float()].\nmedian_sliding_window(Nums, K) ->\n .", "__typename": "CodeSnippetNode" }, { "lang": "Elixir", "langSlug": "elixir", "code": "defmodule Solution do\n @spec median_sliding_window(nums :: [integer], k :: integer) :: [float]\n def median_sliding_window(nums, k) do\n \n end\nend", "__typename": "CodeSnippetNode" } ], "stats": "{\"totalAccepted\": \"136.6K\", \"totalSubmission\": \"344.7K\", \"totalAcceptedRaw\": 136594, \"totalSubmissionRaw\": 344733, \"acRate\": \"39.6%\"}", "hints": [ "The simplest of solutions comes from the basic idea of finding the median given a set of numbers. We know that by definition, a median is the center element (or an average of the two center elements). Given an unsorted list of numbers, how do we find the median element? If you know the answer to this question, can we extend this idea to every sliding window that we come across in the array?", "Is there a better way to do what we are doing in the above hint? Don't you think there is duplication of calculation being done there? Is there some sort of optimization that we can do to achieve the same result? This approach is merely a modification of the basic approach except that it simply reduces duplication of calculations once done.", "The third line of thought is also based on this same idea but achieving the result in a different way. We obviously need the window to be sorted for us to be able to find the median. Is there a data-structure out there that we can use (in one or more quantities) to obtain the median element extremely fast, say O(1) time while having the ability to perform the other operations fairly efficiently as well?" ], "solution": { "id": "86", "canSeeDetail": false, "paidOnly": true, "hasVideoSolution": false, "paidOnlyVideo": true, "__typename": "ArticleNode" }, "status": null, "sampleTestCase": "[1,3,-1,-3,5,3,6,7]\n3", "metaData": "{ \n \"name\": \"medianSlidingWindow\",\n \"params\": [\n { \n \"name\": \"nums\",\n \"type\": \"integer[]\"\n },\n {\n \"name\": \"k\",\n \"type\": \"integer\"\n }\n ],\n \"return\": {\n \"type\": \"double[]\"\n }\n}", "judgerAvailable": true, "judgeType": "large", "mysqlSchemas": [], "enableRunCode": true, "enableTestMode": false, "enableDebugger": true, "envInfo": "{\"cpp\": [\"C++\", \"

Compiled with clang 11 using the latest C++ 20 standard.

\\r\\n\\r\\n

Your code is compiled with level two optimization (-O2). AddressSanitizer is also enabled to help detect out-of-bounds and use-after-free bugs.

\\r\\n\\r\\n

Most standard library headers are already included automatically for your convenience.

\"], \"java\": [\"Java\", \"

OpenJDK 17. Java 8 features such as lambda expressions and stream API can be used.

\\r\\n\\r\\n

Most standard library headers are already included automatically for your convenience.

\\r\\n

Includes Pair class from https://docs.oracle.com/javase/8/javafx/api/javafx/util/Pair.html.

\"], \"python\": [\"Python\", \"

Python 2.7.12.

\\r\\n\\r\\n

Most libraries are already imported automatically for your convenience, such as array, bisect, collections. If you need more libraries, you can import it yourself.

\\r\\n\\r\\n

For Map/TreeMap data structure, you may use sortedcontainers library.

\\r\\n\\r\\n

Note that Python 2.7 will not be maintained past 2020. For the latest Python, please choose Python3 instead.

\"], \"c\": [\"C\", \"

Compiled with gcc 8.2 using the gnu11 standard.

\\r\\n\\r\\n

Your code is compiled with level one optimization (-O1). AddressSanitizer is also enabled to help detect out-of-bounds and use-after-free bugs.

\\r\\n\\r\\n

Most standard library headers are already included automatically for your convenience.

\\r\\n\\r\\n

For hash table operations, you may use uthash. \\\"uthash.h\\\" is included by default. Below are some examples:

\\r\\n\\r\\n

1. Adding an item to a hash.\\r\\n

\\r\\nstruct hash_entry {\\r\\n    int id;            /* we'll use this field as the key */\\r\\n    char name[10];\\r\\n    UT_hash_handle hh; /* makes this structure hashable */\\r\\n};\\r\\n\\r\\nstruct hash_entry *users = NULL;\\r\\n\\r\\nvoid add_user(struct hash_entry *s) {\\r\\n    HASH_ADD_INT(users, id, s);\\r\\n}\\r\\n
\\r\\n

\\r\\n\\r\\n

2. Looking up an item in a hash:\\r\\n

\\r\\nstruct hash_entry *find_user(int user_id) {\\r\\n    struct hash_entry *s;\\r\\n    HASH_FIND_INT(users, &user_id, s);\\r\\n    return s;\\r\\n}\\r\\n
\\r\\n

\\r\\n\\r\\n

3. Deleting an item in a hash:\\r\\n

\\r\\nvoid delete_user(struct hash_entry *user) {\\r\\n    HASH_DEL(users, user);  \\r\\n}\\r\\n
\\r\\n

\"], \"csharp\": [\"C#\", \"

C# 10 with .NET 6 runtime

\"], \"javascript\": [\"JavaScript\", \"

Node.js 16.13.2.

\\r\\n\\r\\n

Your code is run with --harmony flag, enabling new ES6 features.

\\r\\n\\r\\n

lodash.js library is included by default.

\\r\\n\\r\\n

For Priority Queue / Queue data structures, you may use 5.3.0 version of datastructures-js/priority-queue and 4.2.1 version of datastructures-js/queue.

\"], \"ruby\": [\"Ruby\", \"

Ruby 3.1

\\r\\n\\r\\n

Some common data structure implementations are provided in the Algorithms module: https://www.rubydoc.info/github/kanwei/algorithms/Algorithms

\"], \"swift\": [\"Swift\", \"

Swift 5.5.2.

\"], \"golang\": [\"Go\", \"

Go 1.21

\\r\\n

Support https://godoc.org/github.com/emirpasic/gods@v1.18.1 library.

\"], \"python3\": [\"Python3\", \"

Python 3.10.

\\r\\n\\r\\n

Most libraries are already imported automatically for your convenience, such as array, bisect, collections. If you need more libraries, you can import it yourself.

\\r\\n\\r\\n

For Map/TreeMap data structure, you may use sortedcontainers library.

\"], \"scala\": [\"Scala\", \"

Scala 2.13.7.

\"], \"kotlin\": [\"Kotlin\", \"

Kotlin 1.9.0.

\"], \"rust\": [\"Rust\", \"

Rust 1.58.1

\\r\\n\\r\\n

Supports rand v0.6\\u00a0from crates.io

\"], \"php\": [\"PHP\", \"

PHP 8.1.

\\r\\n

With bcmath module

\"], \"typescript\": [\"Typescript\", \"

TypeScript 5.1.6, Node.js 16.13.2.

\\r\\n\\r\\n

Your code is run with --harmony flag, enabling new ES2022 features.

\\r\\n\\r\\n

lodash.js library is included by default.

\"], \"racket\": [\"Racket\", \"

Run with Racket 8.3.

\"], \"erlang\": [\"Erlang\", \"Erlang/OTP 25.0\"], \"elixir\": [\"Elixir\", \"Elixir 1.13.4 with Erlang/OTP 25.0\"], \"dart\": [\"Dart\", \"

Dart 2.17.3

\\r\\n\\r\\n

Your code will be run directly without compiling

\"]}", "libraryUrl": null, "adminUrl": null, "challengeQuestion": null, "__typename": "QuestionNode" } } }