2007 - Find Original Array From Doubled Array (Medium)
Problem Link
https://leetcode.com/problems/find-original-array-from-doubled-array/
Problem Statement
An integer array original
is transformed into a doubled array changed
by appending twice the value of every element in original
, and then randomly shuffling the resulting array.
Given an array changed
, return original
ifchanged
is a doubled array. Ifchanged
is not a doubled array, return an empty array. The elements in original
may be returned in any order.
Example 1:
Input: changed = [1,3,4,2,6,8]
Output: [1,3,4]
Explanation: One possible original array could be [1,3,4]:
- Twice the value of 1 is 1 * 2 = 2.
- Twice the value of 3 is 3 * 2 = 6.
- Twice the value of 4 is 4 * 2 = 8.
Other original arrays could be [4,3,1] or [3,1,4].
Example 2:
Input: changed = [6,3,0,1]
Output: []
Explanation: changed is not a doubled array.
Example 3:
Input: changed = [1]
Output: []
Explanation: changed is not a doubled array.
Constraints:
1 <= changed.length <= 105
0 <= changed[i] <= 105
Approach 1: Sorting + Hashmap
- C++
- Python
- Java
- Go
// Time Complexity: O(N + NlogN)
// Space Complexity: O(N)
// where N is the number of elements in `changed`
class Solution {
public:
// hashmap approach
vector<int> findOriginalArray(vector<int>& changed) {
// if the length of the input is odd, then return {}
// because doubled array must have even length
if (changed.size() & 1) return {};
// count the frequency of each number
unordered_map<int, int> m;
for (auto x: changed) m[x]++;
vector<int> ans;
// sort in ascending order
sort(changed.begin(), changed.end());
// keep the unique elements only in changed
// think of test cases like [0,0,0,0]
// alternatively you can handle it like
// - check if the frequency of 0 is odd, if so, return {}
// - push `0` `m[0] / 2` times to ans
changed.erase(unique(changed.begin(), changed.end()), changed.end());
// so that we can iterate `changed` from smallest to largest
for (auto x : changed) {
// if the number of m[x] is greater than than m[x * 2]
// then there would be some m[x] left
// therefore, return {} here as changed is not a doubled array
if (m[x] > m[x * 2]) return {};
for (int i = 0; i < m[x]; i++) {
// otherwise, we put the element `x` `m[x]` times to ans
ans.push_back(x);
// at the same time we decrease the count of m[x * 2] by 1
// we don't need to decrease m[x] by 1 as we won't use it again
m[x * 2] -= 1;
}
}
return ans;
}
};
// Time Complexity: O(N + KlogK)
// Space Complexity: O(N)
// where N is the number of elements in `changed`
// and K is the number of elements in `uniqueNumbers`
class Solution {
public:
// hashmap approach
vector<int> findOriginalArray(vector<int>& changed) {
// if the length of the input is odd, then return {}
// because doubled array must have even length
if (changed.size() & 1) return {};
// count the frequency of each number
unordered_map<int, int> m;
for (auto x: changed) m[x]++;
vector<int> ans;
vector<int> uniqueNumbers;
// push all unuque numbers to `uniqueNumbers`
for (auto x : m) uniqueNumbers.push_back(x.first);
// sort in ascending order
sort(uniqueNumbers.begin(), uniqueNumbers.end());
// so that we can iterate `changed` from smallest to largest
for (auto x : uniqueNumbers) {
// if the number of m[x] is greater than than m[x * 2]
// then there would be some m[x] left
// therefore, return {} here as changed is not a doubled array
if (m[x] > m[x * 2]) return {};
for (int i = 0; i < m[x]; i++) {
// otherwise, we put the element `x` `m[x]` times to ans
ans.push_back(x);
// at the same time we decrease the count of m[x * 2] by 1
// we don't need to decrease m[x] by 1 as we won't use it again
m[x * 2] -= 1;
}
}
return ans;
}
};
// Time Complexity: O(NlogN)
// Space Complexity: O(N)
// where N is the number of elements in `changed`
class Solution {
public:
// multiset approach
vector<int> findOriginalArray(vector<int>& changed) {
// if the length of the input is odd, then return {}
// because doubled array must have even length
if (changed.size() & 1) return {};
vector<int> ans;
// put all the elements to a multiset
multiset<int> s(changed.begin(), changed.end());
// keep doing the following logic when there is an element in the multiset
while (s.size()) {
// get the smallest element
int smallest = *s.begin();
ans.push_back(smallest);
// remove the smallest element in multiset
s.erase(s.begin());
// if the doubled value of smallest doesn't exist in the multiset
// then return {}
if (s.find(smallest * 2) == s.end()) return {};
// otherwise we can remove its doubled element
else s.erase(s.find(smallest * 2));
}
return ans;
}
};
# Time Complexity: O(NlogN)
# Space Complextiy O(N)
# where N is the number of elements in `changed`
class Solution:
def findOriginalArray(self, changed):
# use Counter to count the frequency of each element in `changed`
cnt, ans = Counter(changed), []
# if the length of the input is odd, then return []
# because doubled array must have even length
if len(changed) % 2: return []
# sort in ascending order
for x in sorted(cnt.keys()):
# if the number of cnt[x] is greater than than cnt[x * 2]
# then there would be some cnt[x] left
# therefore, return [] here as changed is not a doubled array
if cnt[x] > cnt[x * 2]: return []
# handle cases like [0,0,0,0]
if x == 0:
# similarly, odd length -> return []
if cnt[x] % 2:
return []
else:
# add `0` `cnt[x] // 2` times
ans += [0] * (cnt[x] // 2)
else:
# otherwise, we put the element `x` `cnt[x]` times to ans
ans += [x] * cnt[x]
cnt[2 * x] -= cnt[x]
return ans
class Solution {
// counting sort approach
public int[] findOriginalArray(int[] changed) {
int n = changed.length, j = 0;
// if the length of the input is odd, then return []
// because doubled array must have even length
if (n % 2 == 1) return new int[]{};
int[] ans = new int[n / 2];
// alternatively, you can find the max number in `changed`
// then use new int[2 * mx + 1]
int[] cnt = new int[200005];
// count the frequency of each number
for (int x : changed) cnt[x] += 1;
// iterate from 0 to max number
for (int i = 0; i < 200005; i++) {
// check if the count of number i is greater than 0
if (cnt[i] > 0) {
// number i exists, decrease by 1
cnt[i] -= 1;
// look for the doubled value
if (cnt[i * 2] > 0) {
// doubled value exists, decrease by 1
cnt[i * 2] -= 1;
// add this number to ans
ans[j++] = i--;
} else {
// cannot pair up, return []
return new int[]{};
}
}
}
return ans;
}
}
// 1. Mapping the numbers with repetitions.
// 2. Sort them in ascending order
// 3. Subtract the doubled from the maps and make the original item 0.
// 4. Checking if everything is 0.
func findOriginalArray(changed []int) []int {
// As every number has double, if it's odd, directly return empty.
isOdd := len(changed) & 1
if(isOdd == 1){
return []int{}
}
// Converting everything to a map, counting the repetitions.
changedMap := make(map[int]int)
for _, elem := range changed {
if _, ok:= changedMap[elem]; ok {
changedMap[elem] = changedMap[elem] +1
}else {
changedMap[elem] = 1;
}
}
// Creating a unique set of all the repetitions.
// Sorting the set/unique array in ascending order as we can easily remove the elements.
var set []int
for key := range changedMap {
set = append(set, key)
}
sort.Ints(set)
var original []int;
for _, elem := range set {
if _, ok:= changedMap[elem*2]; ok {
x := changedMap[elem]
y := changedMap[elem*2]
if(elem == 0){ // In the case of [0,0,0,0] the array has to be [0,0]
x = x/2
}
// We append by the repetitions.
for i := 0; i < x; i++ {
original = append(original, elem)
}
// In the case of [2,1,2,4,2,4] we have 2s more than 1s. so we subtract 2s from 1s.
changedMap[elem*2] = y - x;
changedMap[elem] = 0;
}
}
// Only if all the changedMap has 0 elements in them we make we consider they are all doubled.
for _, elem := range changedMap {
if (elem != 0){
return []int{}
}
}
return original
}