Morgana is learning computer vision, and he likes cats, too. One day he wants to find the cat movement from a cat video. To do this, he extracts cat features in each frame. A cat feature is a two-dimension vector <xx, yy>. If x_ixi​ = x_jxj​and y_iyi​ = y_jyj​, then <x_ixi​, y_iyi​> <x_jxj​, y_jyj​> are same features.

So if cat features are moving, we can think the cat is moving. If feature <aa, bb> is appeared in continuous frames, it will form features movement. For example, feature <aa , bb > is appeared in frame 2,3,4,7,82,3,4,7,8, then it forms two features movement 2-3-42−3−4 and 7-87−8 .

Now given the features in each frames, the number of features may be different, Morgana wants to find the longest features movement.

Input

First line contains one integer T(1 \le T \le 10)T(1≤T≤10), giving the test cases.

Then the first line of each cases contains one integer nn (number of frames),

In The next nn lines, each line contains one integer k_iki​ ( the number of features) and 2k_i2ki​intergers describe k_iki​ features in ith frame.(The first two integers describe the first feature, the 33rd and 44th integer describe the second feature, and so on).

In each test case the sum number of features NNwill satisfy N \le 100000N≤100000 .

Output

For each cases, output one line with one integers represents the longest length of features movement.

样例输入复制

1
8
2 1 1 2 2
2 1 1 1 4
2 1 1 2 2
2 2 2 1 4
0
0
1 1 1
1 1 1

样例输出复制

3

题目来源

ACM-ICPC 2018 徐州赛区网络预赛

 
 
题意:n个集合,每个集合有ki个点,问连续在多个集合中出现的次数最多的点的次数
分析:用一个map存下每个点的最后一次出现位置和连续出现次数,连续出现次数通过最后一次出现位置来判断,每次更新连续出现次数时取下最大值
AC代码:
#include <map>
#include <set>
#include <stack>
#include <cmath>
#include <queue>
#include <cstdio>
#include <vector>
#include <string>
#include <bitset>
#include <cstring>
#include <iomanip>
#include <iostream>
#include <algorithm>
#define ls (r<<1)
#define rs (r<<1|1)
#define debug(a) cout << #a << " " << a << endl
using namespace std;
typedef long long ll;
const ll maxn = 1e5+10;
const ll mod = 2e9+7;
const double pi = acos(-1.0);
const double eps = 1e-8;
map<ll,pair<ll,ll> >mp;
map<ll,ll> mm;
int main() {
ll k, n, a, b, T;
scanf("%lld",&T);
while( T -- ) {
scanf("%lld",&n);
ll maxa = 0;
mp.clear();
for( ll i = 1; i <= n; i ++ ) {
scanf("%lld",&k);
mm.clear();
while( k -- ) {
scanf("%lld%lld",&a,&b);
ll t = a*mod + b; //通过乘mod将每个坐标转化成一个不同的值
if( mp[t].first == i-1 ) {
mp[t].second = mp[t].second + 1, mp[t].first = i;
} else if( mm[t] ) {
continue;
} else {
mp[t].second = 1, mp[t].first = i;
}
mm[t] ++;
maxa = max(maxa,mp[t].second);
}
}
if( maxa == 1 ) {
maxa = 0;
}
printf("%lld\n",maxa);
}
return 0;
}

  

Features Track 2018徐州icpc网络赛 思维的更多相关文章

  1. Trace 2018徐州icpc网络赛 思维+二分

    There's a beach in the first quadrant. And from time to time, there are sea waves. A wave ( xx , yy) ...

  2. Ryuji doesn't want to study 2018徐州icpc网络赛 树状数组

    Ryuji is not a good student, and he doesn't want to study. But there are n books he should learn, ea ...

  3. Trace 2018徐州icpc网络赛 (二分)(树状数组)

    Trace There's a beach in the first quadrant. And from time to time, there are sea waves. A wave ( xx ...

  4. ICPC 2018 徐州赛区网络赛

    ACM-ICPC 2018 徐州赛区网络赛  去年博客记录过这场比赛经历:该死的水题  一年过去了,不被水题卡了,但难题也没多做几道.水平微微有点长进.     D. Easy Math 题意:   ...

  5. ACM-ICPC 2018 徐州赛区(网络赛)

    目录 A. Hard to prepare B.BE, GE or NE F.Features Track G.Trace H.Ryuji doesn't want to study I.Charac ...

  6. Supreme Number 2018沈阳icpc网络赛 找规律

    A prime number (or a prime) is a natural number greater than 11 that cannot be formed by multiplying ...

  7. 2019 徐州icpc网络赛 E. XKC's basketball team

    题库链接: https://nanti.jisuanke.com/t/41387 题目大意 给定n个数,与一个数m,求ai右边最后一个至少比ai大m的数与这个数之间有多少个数 思路 对于每一个数,利用 ...

  8. ACM-ICPC 2018 徐州赛区网络预赛 F. Features Track

    262144K   Morgana is learning computer vision, and he likes cats, too. One day he wants to find the ...

  9. ACM-ICPC 2018 徐州赛区网络预赛 G. Trace (思维,贪心)

    ACM-ICPC 2018 徐州赛区网络预赛 G. Trace (思维,贪心) Trace 问答问题反馈 只看题面 35.78% 1000ms 262144K There's a beach in t ...

随机推荐

  1. 【Java】Map

    今天用到了键-值对,于是想起了 Java 的 Map,由于之前并不很熟悉,就看了下源码,如下: /* * Copyright (c) 1997, 2006, Oracle and/or its aff ...

  2. MOCTF-MISC-writeup

    小菜鸡终于想要开通博客,要开始写东西了.第一次,献给了MOCTF MISC writeup,各种借鉴大哥们的writeup,写的不好的话,算了,后果自负(嘤嘤嘤). MOCTF平台地址:www.moc ...

  3. 整理用Java实现数字转化成字符串左边自动补零方法

    Java 中给数字左边补0 (1)方法一 import java.text.NumberFormat; public class NumberFormatTest { public static vo ...

  4. Hadoop 系列(一)—— 分布式文件系统 HDFS

    一.介绍 HDFS (Hadoop Distributed File System)是 Hadoop 下的分布式文件系统,具有高容错.高吞吐量等特性,可以部署在低成本的硬件上. 二.HDFS 设计原理 ...

  5. oracle 正确删除归档日志,并清除 V$ARCHIVED_LOG 数据

    1. 连接 RMAN 管理 rman target / 2. 查看归档日志列表 RMAN> crosscheck archivelog all; 3. 删除所有归档日志 RMAN> DEL ...

  6. 0x15 字符串

    KMP算法 next数组的求法 void calc_next() { next[]=; , j=; i<=n; ++i) { &&a[i]!=a[j+]) j=next[j]; ...

  7. ccf 201903-5 317号子任务(60分)

    看到这题,第一印象,用dijkstra算法求n次单源最短路,时间复杂度O(n^3),超时30分妥妥的. 于是用优先队列优化,O(n*mlogm),快很多,但依然30. 那么不妨换一种思路,题目要求的是 ...

  8. 搞定java String校招面试题

    今天大致的阅读了String类的源码,并刷了常见的面试题,在此做个笔记. 面试题一:判断下列程序运行结果 package String_test; public class test_1 { publ ...

  9. Flink 源码解析 —— Flink JobManager 有什么作用?

    JobManager 的作用 https://t.zsxq.com/2VRrbuf 博客 1.Flink 从0到1学习 -- Apache Flink 介绍 2.Flink 从0到1学习 -- Mac ...

  10. 工作中常见的五种技术leader

    力不从心型 在工作中有种技术leader,总认为自己是最好的.在方案设计的时候,自己有一种方案,下属有一种方案.leader非要别人听他的.如果两种方案没有优劣之分,比较建议的做法是让真正实施的人按照 ...