Skip to main content

Valid Triangle Number

Problem Description

Given an integer array nums, return the number of triplets chosen from the array that can make triangles if we take them as side lengths of a triangle.

Examples

Example 1:

Input: nums = [2,2,3,4]
Output: 3
Explanation: Valid combinations are:
2,3,4 (using the first 2)
2,3,4 (using the second 2)
2,2,3

Example 2:

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

Constraints

  • 1 <= nums.length <= 1000
  • 0 <= nums[i] <= 1000

Solution for Valid Triangle Number

Approach 1 Brute Force [Time Limit Exceeded]

The condition for the triplets (a,b,c) representing the lengths of the sides of a triangle, to form a valid triangle, is that the sum of any two sides should always be greater than the third side alone. i.e. a+b>c, b+c>a, a+c>b.

The simplest method to check this is to consider every possible triplet in the given nums array and checking if the triplet satisfies the three inequalities mentioned above. Thus, we can keep a track of the count of the number of triplets satisfying these inequalities. When all the triplets have been considered, the count gives the required result.

Caution: The brute force approach is included here because it is an intuitive way to approach this problem. However, when there are 10310^3 numbers, the if statement will be checked approximately 10910^9 times. Thus, this approach will result in TLE. In the following approaches we will discuss ways to optimize our solution.

Code in Different Languages

Written by @Shreyash3087
#include <vector>

class Solution {
public:
int triangleNumber(std::vector<int>& nums) {
int count = 0;
for (size_t i = 0; i < nums.size() - 2; i++) {
for (size_t j = i + 1; j < nums.size() - 1; j++) {
for (size_t k = j + 1; k < nums.size(); k++) {
if (nums[i] + nums[j] > nums[k] &&
nums[i] + nums[k] > nums[j] &&
nums[j] + nums[k] > nums[i]) {
count++;
}
}
}
}
return count;
}
};

Complexity Analysis

Time Complexity: O(N3)O(N^3)

Reason: Three nested loops are there to check every triplet.

Space Complexity: O(1)O(1)

Reason: Constant space is used.

Algorithm

If we sort the given nums array once, we can solve the problem more efficiently. This is because if we consider a triplet (a, b, c) such that abca \leq b \leq c, we need not check all three inequalities for the validity of the triangle formed by them. Only the condition a+b>ca + b > c is necessary. This happens because cbc \geq b and cac \geq a. Thus, adding any number to c will always produce a sum greater than either a or b considered alone. Therefore, the inequalities c+a>bc + a > b and c+b>ac + b > a are satisfied implicitly by the property a<b<ca < b < c.

From this, we get the idea that we can sort the given nums array. Then, for every pair (nums[i], nums[j]) considered from the beginning of the array, such that j>ij > i (leading to nums[j]nums[i]nums[j] \geq nums[i], we can find the count of elements nums[k] (where k>jk > j), which satisfy the inequality nums[k]>nums[i]+nums[j]nums[k] > nums[i] + nums[j]. We can do this for every pair (i, j) considered to get the required result.

We can also observe that, since we've sorted the nums array, as we traverse towards the right for choosing the index k (for number nums[k]), the value of nums[k] could increase or remain the same (doesn't decrease relative to the previous value). Thus, there will exist a right limit on the value of index k, such that the elements satisfy nums[k]>nums[i]+nums[j]nums[k] > nums[i] + nums[j]. Any elements beyond this value of k won't satisfy this inequality as well, which is obvious.

Thus, if we can find this right limit value of k (indicating the element just greater than nums[i]+nums[j]nums[i] + nums[j]), we can conclude that all the elements in the nums array in the range (j+1,k1)(j + 1, k - 1) (both included) satisfy the required inequality. Thus, the count of elements satisfying the inequality will be given by kj1k - j - 1.

Since the nums array has been sorted, we can use Binary Search to find this right limit of k.

Another point to be observed is that once we find a right limit index k(i,j)k_{(i,j)} for a particular pair (i,j) chosen, when we choose a higher value of j for the same value of i, we need not start searching for the right limit k(i,j+1)k_{(i,j+1)} from the index j+2j. Instead, we can start off from the index k(i,j)k_{(i,j)} directly where we left off for the last j chosen.

This holds correct because when we choose a higher value of j(higher or equal nums[j] than the previous one), all the nums[k], such that k<k(i,j)k < k_{(i,j)} will obviously satisfy nums[i]+nums[j]>nums[k] for the new value of j chosen.

By taking advantage of this observation, we can limit the range of Binary Search for k to shorter values for increasing values of j considered while choosing the pairs (i,j).

Code in Different Languages

Written by @Shreyash3087
public class Solution {
int binarySearch(int nums[], int l, int r, int x) {
while (r >= l && r < nums.length) {
int mid = (l + r) / 2;
if (nums[mid] >= x)
r = mid - 1;
else
l = mid + 1;
}
return l;
}
public int triangleNumber(int[] nums) {
int count = 0;
Arrays.sort(nums);
for (int i = 0; i < nums.length - 2; i++) {
int k = i + 2;
for (int j = i + 1; j < nums.length - 1 && nums[i] != 0; j++) {
k = binarySearch(nums, k, nums.length - 1, nums[i] + nums[j]);
count += k - j - 1;
}
}
return count;
}
}

Complexity Analysis

Time Complexity: O(n2logn)O(n^2logn)

Reason: In worst case inner loop will take nlog⁡n (binary search applied N times).

Space Complexity: O(logn)O(logn)

Reason: Sorting takes O(log⁡n) space.

References


Authors:

Loading...