心情不爽，写个top K 解闷

心情不爽

写个程序发泄一下

前几天看到师弟们在找工作，

想起自己找工作的时候常被问道一个问题：

给定10个有序文件，每个100M，假设只包含数据，要取出这10*100M数据里面的前100个

这个是典型的top K问题，虽然我知道有很多种方法可以解，

但惭愧的是从来没有自己实现过这个代码，面试笔试也就这么糊弄过去了

今天用多路归并+堆排序实现了下

代码如下：（凑合着看吧，在vc测试通过）

// heapSort.cpp : Defines the entry point for the console application.
//

#include<iostream>
#include<vector>
#include<list>
using namespace std;

//#define PARENT(i) (i/2)
#define LEFT(i) (2*i+1)
//#define RIGHT(i) (2*i+1) 
struct Node { 
    int value; 
    Node *next; 
};

void min_heapity(Node* Array,int i,int n)
{
	int l = LEFT(i);
	int r = l+1;//RIGHT(i)
	int minest = i;
	if (l<=n-1 && Array[l].value < Array[i].value)//n-1是heap-size
		minest = l;
	else
		minest = i;
	if (r<=n-1 && Array[r].value < Array[minest].value)
	{
		minest = r;
	}
	if (minest != i)
	{
		Node temp = Array[minest];
		Array[minest] = Array[i];
		Array[i] = temp;
		min_heapity(Array,minest,n);//递归
	}

}//建堆
void bulid_heap(Node heap[],int n)
{
	
	for(int i = n/2-1;i>=0;i--)//注意从n/2开始，到0结束，逆向的
	{
		min_heapity(heap,i,n);
	}

}

//堆排序算法
void heap_sort(Node* Array,int n)
{
	
	bulid_heap(Array,n);
	int tmp_n = n;
	for (int i = n-1;i>=1;i-- )
	{
		Node temp = Array[i];
		Array[i] = Array[0];
		Array[0] = temp;
		min_heapity(Array,0,--tmp_n);
	}


}

void kMerge(vector<vector<Node> > &kLists) { 
        int heapSize = kLists.size(); 
        //Node heap[] = new Node[heapSize + 1]; 
		Node *heap = (Node *)malloc(sizeof(Node)*heapSize); 
		// 取出K路中每个vector的第一个元素建堆。 
        for (int i = 0; i < heapSize; i++) 
        {    heap[i].value = kLists.at(i).at(0).value;
			 heap[i].next = kLists.at(i).at(0).next;
		}
        
        bulid_heap(heap, heapSize); 
		for (i = 0;i<heapSize;i++)
		{
			cout<<"build heap "<<heap[i].value<<endl;
		}
        vector<Node> resultList; 
        while (heapSize > 0) { 
 
             Node minNode = heap[0];
            resultList.push_back(minNode);
  
            if (minNode.next == NULL) { 
                heap[0] = heap[heapSize-1]; 
                heapSize -= 1; 
            } else { 
                heap[0].value = minNode.next->value;
				heap[0].next = minNode.next->next;
            } 
 
            min_heapity(heap, 0, heapSize); 
			for (i = 0;i<heapSize;i++)
			{
				cout<<"min heap "<<heap[i].value<<endl;
			}
			cout<<endl;
        } 
		for (vector<Node>::iterator t = resultList.begin(); t != resultList.end();++t)
		{
			cout<<"test "<<(*t).value<<endl;
		}
        //return resultList; 
    } 
void creatList(int *a,vector<Node> &v,int length)
{
	
	Node *preNode = NULL; 
	for (int i=0;i<length; i++)
	{
		v.at(i).value = a[i];
		v.at(i).next = NULL;
		if (preNode != NULL) 
			preNode->next = &(v.at(i)); 
        preNode = &(v.at(i));
	}
	//return v;
}
int main()
{
	int a1[] = {2, 5, 8, 10};
	vector<Node> vector1;
	vector1.resize(4);
	creatList(a1, vector1, 4);

	int a2[] = {3, 4, 14};
	vector<Node> vector2;
	vector2.resize(3);
	creatList(a2, vector2, 3);

	int a3[] = {9, 11, 13, 15, 18};
	vector<Node> vector3;
	vector3.resize(5);
	creatList(a3, vector3, 5);
	
	int a4[] = {7, 12, 17, 22, 25};
	vector<Node> vector4;
	vector4.resize(5);
	creatList(a4, vector4, 5);
	
	int a5[] = {6, 16, 19, 21};
	vector<Node> vector5;
	vector5.resize(4);
	creatList(a5, vector5, 4);

	vector<vector<Node> > Kvec;
	Kvec.push_back(vector1);
	Kvec.push_back(vector2);
	Kvec.push_back(vector3);
	Kvec.push_back(vector4);
	Kvec.push_back(vector5);
	kMerge(Kvec);
	for(vector<vector<Node> >::iterator t = Kvec.begin() ;t != Kvec.end(); t++ )
	{
		for(vector<Node>::iterator pp = t->begin(); pp!= t->end() ; pp++ )
			cout<<(*pp).value<<endl;
	}
	

	return 0;

}

算是自己在csdn的第一篇blog，mark下吧

不知道自己能不能够坚持写blog

毕竟现在自己的工作和编程关系是在不大

让我放弃C++，真的挺不甘心的