- Find the Duplicate Number
Given an array nums containing n + 1 integers where each integer is between 1 and n (inclusive), prove that at least one duplicate number must exist. Assume that there is only one duplicate number, find the duplicate one.
Example
Given nums = [5,5,4,3,2,1] return 5
Given nums = [5,4,4,3,2,1] return 4
Notice
You must not modify the array (assume the array is read only).
You must use only constant, O(1) extra space.
Your runtime complexity should be less than O(n^2).
There is only one duplicate number in the array, but it could be repeated more than once.
这题看起来很容易,其实很难。首先这题不能排序,也不能用hash,因为题目要求不能动数组,只能用O(1)空间。也不用考虑XOR了,因为重复数可以重复n多次(n>1)。
思路1:基于binary search。
假设数组[6,6,7,6,3,6,6,6,6,6,10],共11个元素。n=10。假设重复数是d,
规律1:对于某个数x,如果数组中小于等于x的数的个数<x,那么x<d。
举例: 数组中<=3的数的个数是1。原因很简单,因为<3的数中某些被6替代了。
规律2:对于某个数x,如果数组中小于等于x的数的个数>x,那么x>=d。
举例: 数组中<=6的数的个数是9,<=7的个数是10,<=10的个数是11。
注意这里是x>=d,而不是x>d。
规律3:对于某个数x,如果数组中小于等于x的数的个数=x,那么x<d。
举例:数组[5,5,4,3,2,1], 小于等于3的个数是3,而3<5。
实际上,规律1和规律3可以合并,即
规律4:对于某个数x,如果数组中小于等于x的数的个数<=x,那么x<d。
注意:
- calCount()里面比较大小一定要用<=,而不是<
- 该模板最后结束while()循环后,把范围缩小到start和end两个元素。如果calCount(start) > start, 根据规律2,start >= dup,既然start已经大于dup,那当然轮不到end了。否则calCount<=start,根据规律4,start < dup,此时返回end即可。
代码如下:
class Solution {
public:
/**
* @param nums: an array containing n + 1 integers which is between 1 and n
* @return: the duplicate one
*/
int findDuplicate(vector<int> &nums) {
int len = nums.size();
int start = 1, mid = 0, end = len - 1;
while (start + 1 < end) {
mid = start + (end - start) / 2;
if (calCount(nums, mid) > mid) { // it shows mid >= dup
end = mid;
} else {
start = mid;
}
}
if (calCount(nums, start) > start) return start;
return end;
}
private:
int calCount(vector<int> &nums, int n) {
int count = 0;
int len = nums.size();
for (int i = 0; i < len; ++i) {
if (nums[i] <= n) count++; //it should be <= here! not <
}
return count;
}
};
注意:如果在while循环中,我们要以calCount<mid来判断,则应写成
int findDuplicate(vector<int> &nums) {
int len = nums.size();
int start = 1, mid = 0, end = len - 1;
while (start + 1 < end) {
mid = start + (end - start) / 2;
if (calCount(nums, mid) <= mid) { //注意这里是<=, shows mid<=dup
start = mid;
} else {
end = mid;
}
}
if (calCount(nums, start) > start) return start;
return end;
}
为什么这里要用 <=呢? 看看上面的规律4就明白了,因为<和==是同一种情况。为什么上面的用>呢?看看规律2就明白了,因为>是单独一种情况。
二刷:
class Solution {
public:
/**
* @param nums: an array containing n + 1 integers which is between 1 and n
* @return: the duplicate one
*/
int findDuplicate(vector<int> &nums) {
int start = 1, end = nums.size() - 1; //注意这里的start, end代表值,不是索引, start必须是1,不是0。因为nums没有排序
while (start + 1 < end) {
int mid = start + (end - start) / 2;
if (count(nums, mid) <= mid) {
start = mid;
} else {
end = mid;
}
}
//#if 0
if (count(nums, start) <= start) { //注意:我们要找第一个count(nums, x)>=x的数,
return end;
}
return start;
//#endif
#if 0
if (count(nums, start) > start) return start; //这个也可以
return end;
#endif
}
private:
int count(vector<int> &nums, int target) {
int res = 0;
for (auto num : nums) {
if (num <= target) res++;
}
return res;
}
};
思路2:利用环状链表,时间复杂度可以达到O(n)。下次再研究。
参考链接:
http://bookshadow.com/weblog/2015/09/28/leetcode-find-duplicate-number/
http://www.cnblogs.com/grandyang/p/4843654.html
思路3:位运算,时间复杂度也是O(n),下次再研究。
参考链接:
http://www.cnblogs.com/grandyang/p/4843654.html
解法3:每次把nums[i]当index,把对应的数乘以-1,如果已经是负数了,说明是duplicate。
class Solution {
public:
/**
* @param nums: an array containing n + 1 integers which is between 1 and n
* @return: the duplicate one
*/
int findDuplicate(vector<int> &nums) {
int n = nums.size();
for (int i = 0; i < n; i++) {
if (nums[abs(nums[i]) - 1] < 0) {
return abs(nums[i]);
} else {
nums[abs(nums[i]) - 1] = -1 * nums[abs(nums[i]) - 1];
}
}
return -1;
}
};