This problem equals to finding the Kth maximum element in two sorted arrays. In order to reach the complexity of O(log(m+n)), we can use binary search recursively.
Compare the K/2th elements of A and B,
If A[k/2]<B[k/2], throw away the smaller part of A. Notice this won't affect the Kth element!
If there's less than k/2 elements in array, simply throw away this array. And the Kth element must be in another array!
class Solution {
public:
double findMedianSortedArrays(int A[], int m, int B[], int n) {
int s = m+n;
if(s%2) return findKth(A,m,B,n,s/2+1);
else return 0.5*(findKth(A,m,B,n,s/2)+findKth(A,m,B,n,s/2+1));
}
int findKth(int A[], int m, int B[], int n, int k) {
if(m>n) return findKth(B,n,A,m,k);
if(m==0) return B[k-1];
if(k==1) return min(A[0],B[0]);
int a = min(m,k/2);
int b = k-a;
if(A[a-1]<B[b-1]) return findKth(A+a,m-a,B,n,k-a);
if(A[a-1]>B[b-1]) return findKth(A,m,B+b,n-b,k-b);
return A[a-1];
}
};