NNS - Nearest Neighbor Search
You are given N (N <= 100 000) d-dimensional (1 <= d <= 5) points, each having integer coordinates (X1, X2, ..., Xd). Then Q (Q <= 100 000) queries follow. For each query you are given a d-dimensional point (not necessarily from the given N) and you are to find the squared Euclidean distance to the closest point from the given N.
The coordinates of all N+Q points are generated randomly between -1 000 000 000 and 1 000 000 000.
The squared Euclidean distance between two points A and B is the sum of (A.Xi-B.Xi)*(A.Xi-B.Xi) for i=1, 2, ..., d.
Input
The first line contains three numbers - N, d and Q.
The next N lines contain d integers each - the coordinates of a point.
The next Q lines contain d integers each - the coordinates of a query point.
Output
Print the answer for each of the Q queries on separate lines.
Example
Input: 2 2 2 1 1 2 2 1 1 3 3 Output: 0 2
hide comments
visleck:
2017-08-29 11:33:25
@extazy: Constraints on d??
|
Added by: | Extazy |
Date: | 2017-08-29 |
Time limit: | 2s |
Source limit: | 50000B |
Memory limit: | 1536MB |
Cluster: | Cube (Intel G860) |
Languages: | All |