一些排序算法

 

鸽巢排序，排序字节串、宽字节串最快的排序算法，计数排序的变种（将计数缓冲区大小固定，少一次遍历开销），速度是STL中std::sort的20多倍，更重要的是实现极其简单！缺点是需要一个size至少等于待排序数组取值范围的缓冲区，不适合int等大范围数据

C/C++ code

void PigeonholeSort(BYTE *array, int length)
{
    int b[256] = {0};
    int i,k,j = 0;
    for(i=0; i<length; i++)
        b[array[i]]++;
    for(i=0; i<256; i++)
        for(k=0; k<b[i]; k++)
            array[j++] = i;
}

多一次遍历的计数排序，排序字节串的话速度约是鸽巢排序的一半

C/C++ code

void CountingSort(BYTE *array, int length)
{
    int t;
    int i, z = 0;
    BYTE min,max;
    int *count;
    min = max = array[0];
    for(i=0; i<length; i++)
    {
        if(array[i] < min)
            min = array[i];
        else if(array[i] > max)
            max = array[i];
    }
    count = (int*)malloc((max-min+1)*sizeof(int));
    for(i=0; i<max-min+1; i++)
        count[i] = 0;
    for(i = 0; i < length; i++)
        count[array[i]-min]++;

    for(t = min; t <= max; t++)
        for(i = 0; i < count[t-min]; i++)
            array[z++] = (BYTE)t;
    free(count);
}

快速排序，快排最标准的递归实现，速度约是std::sort的一半

C/C++ code

void swap(BYTE *a,BYTE *b)
{
    BYTE tmp;
    if ( a != b )
    {
        tmp = *a;
        *a = *b;
        *b = tmp;
    }
}

int partition(BYTE *arr,int left, int right)
{
    int i = left - 1, j = right;
    BYTE v = arr[right];
    while(1)
    {
        while(arr[++i] < v);
        while(arr[--j] > v)
            if(j == 1)
                break;
        if(i >= j)
            break;
        swap(&arr[i],&arr[j]);
    }
    swap(&arr[i],&arr[right]);
    return i;
}

void quicksort(BYTE *arr, int left, int right)
{
    if (left < right)
    {
        int i = partition(arr,left,right);

        quicksort(arr,left,i-1);
        quicksort(arr,i+1,right);
    }
}

void QuickSort(BYTE *array,int length)
{
    quicksort(array,0,length-1);
}

这是速度与std::sort相当的三路划分快排

C/C++ code

void swap(BYTE *a,BYTE *b)
{
    BYTE tmp;
    if ( a != b )
    {
        tmp = *a;
        *a = *b;
        *b = tmp;
    }
}

void quicksort(BYTE *arr, int left, int right)
{
    if (left < right)
    {
        BYTE v = arr[right];
        int i = left - 1,j = right,p = left - 1,q = right,k=0;
        while (1)
        {
            while (arr[++i] < v);
            while (arr[--j] > v)
                if(j==left)
                    break;
            if (i >= j)
                break;
            swap(&arr[i], &arr[j]);
            if(arr[i] == v)
            {
                p++;
                swap(&arr[p],&arr[i]);
            }
            if(arr[j] == v)
            {
                q--;
                swap(&arr[q],&arr[j]);
            }
        }
        swap(&arr[i],&arr[right]);
        j = i - 1;
        i++;
        for(k=left; k<=p; k++,j--)
            swap(&arr[k],&arr[j]);
        for(k=right-1; k>=q; k--,i++)
            swap(&arr[k],&arr[i]);
        quicksort(arr,left,j);
        quicksort(arr,i,right);
    }
}

void QuickSort(BYTE *array,int length)
{
    quicksort(array,0,length-1);
}

相当简单的梳排序，效率是std::sort的三分之一

C/C++ code

void CombSort(BYTE *arr, int size)
{
    UINT gap = size, swapped = 1, i = 0;
    BYTE swap = 0;
    while ((gap > 1) || swapped)
    {
        if (gap > 1)
            gap = gap / 1.3;
        swapped = 0; 
        i = 0;
        while ((gap + i) < size)
        {
            if (arr[i] - arr[i + gap] > 0)
            {
                swap = arr[i];
                arr[i] = arr[i + gap];
                arr[i + gap] = swap;
                swapped = 1;
            }
            ++i;
        }
    }
}

LSD基数排序，与std::sort速度相当，但是需要一个与输入缓冲一样大的缓冲区

C/C++ code

#define R 256

#define digit(a, d)    ( a >> 8*d )

static BYTE *aux;

void radix_sort(BYTE *arr, int left, int right)
{
    if(left < right)
    {
        int d = 0;
        for(d=3; d>=0; d--)
        {
            int i=0, j=0, count[R+1];
            for(j=0; j<R; j++)
                count[j] = 0;
            for(i=left; i<=right; i++)
                count[digit(arr[i],d) + 1]++;
            for(j=1; j<R; j++)
                count[j] += count[j-1];
            for(i=left; i<=right; i++)
                aux[count[digit(arr[i],d)]++] = arr[i];
            for(i=left; i<=right; i++)
                arr[i] = aux[i-1];
        }    
    }
}

void RadixSort(BYTE *array,int length)
{
    aux = (BYTE*)malloc(length);
    radix_sort(array,0,length-1);
    free(aux);
}

归并排序，效率越是std::sort的六分之一，通常的实现是递归，但和快排不同，归并改循环极其容易

C/C++ code

void merge(BYTE *array, int low, int mid, int high)
{
    int i, k;
    BYTE *temp = (BYTE *) malloc(high-low+1);
    int begin1 = low;
    int end1 = mid;
    int begin2 = mid + 1;
    int end2 = high;

    for (k = 0; begin1 <= end1 && begin2 <= end2; ++k)
        if(array[begin1]<array[begin2])
            temp[k] = array[begin1++];
        else
            temp[k] = array[begin2++];    
    while(begin1<=end1)
        temp[k++] = array[begin1++];
    while(begin2<=end2)
        temp[k++] = array[begin2++];
    for (i = 0; i < (high-low+1); i++)
        array[low+i] = temp[i];
    free(temp);
}

void merge_sort(BYTE *array, UINT first, UINT last)
{
    UINT mid,i;
    for(mid=1; mid<=last-first; mid += mid)
        for(i=first; i<=last-mid; i+=mid+mid)
            merge(array,i,i+mid-1,min(i+mid+mid-1,last));
}

void MergeSort(BYTE *array, UINT length)
{
    merge_sort(array,0,length-1);
}

这是堆排序，相对复杂些，效率是std::sort的四分之一

C/C++ code

UINT parent(UINT i)
{
    return (UINT)floor(i / 2);
}

UINT left(UINT i)
{
    return 2 * i;
}

UINT right(UINT i)
{
    return (2 * i + 1);
}

void Max_Heapify(BYTE *A, UINT i, UINT length)
{
    UINT l = left(i);
    UINT r = right(i);
    UINT largest;
    BYTE temp;
    if(l < length && A[l] > A[i])
        largest = l;
    else
        largest = i;
    if(r < length && A[r] > A[largest])
        largest = r;
    if(largest != i)
    {
        temp = A[i];
        A[i] = A[largest];
        A[largest] = temp;
        Max_Heapify(A, largest, length);
    }
}

void Build_Max_Heap(BYTE *A, UINT length)
{
    int i = 0;
    for(i = length; i >= 0; i--)
        Max_Heapify(A, i, length);
}

void HeapSort(BYTE *A, UINT length)
{
    BYTE temp;
    int i = 0;
    Build_Max_Heap(A,length);
    for(i = length - 1; i >= 1; i--)
    {
        temp = A[0];
        A[0] = A[i];
        A[i] = temp;
        length = length - 1;
        Max_Heapify(A, 0, length);
    }
}

至于冒泡和选择，和以上这几种效率相差几个数量级