数据结构与算法-大数据排序

练习1;
    0-1999的随机数1000个,已知无重复,求最快排序算法;
    答案:
        利用数组下标;
        1-> 定义一个长度是2000的数组arr,清0;
        2-> 遍历1000个随机数,arr[随机数] = 1;
        3-> 打印值是1的数组下标就ok;
练习2:
    一个文件中包含n个正整数,其中n=10的7次方;已知数据不重复;
    请编写程序,将文件中的数字按升序排序;
    根据练习1的思路,但是为节省空间,每个数只用1个bit;
    #define NUM 10000000
    申请的数组为a[1 + NUM/32];
参考代码如下

/*
 * generate.c
 */
/* Include Files */
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <errno.h>
#include <unistd.h>
#include <time.h>
#include <stdlib.h>

/* MACROS */
#define    NUM        10000000
#define    SHIFT    5
#define MASK    0x1F

/* CONSTANTS AND VARIABLES */
int a[1 + NUM/32] = { 0 }; /* define an arry and initialize every member */

/* INTERNAL FUNCTIONS */
void setbit(int n) {
    /* set flag */
    a[n >> SHIFT] |= 1 << (n & MASK); /* a[n/32] |= 1<<(n%32) */
}
void cl_bit(int n) {
    /* clear flag */
    a[n >> SHIFT] &= ~(1 << (n & MASK)); /* a[n/32] |= 1<<(n%32) */
}
int testbit(int n) {
    /* if the n bit exist, resturn 1 */
    return a[n >> SHIFT] & (1 << ( n & MASK));
}
int test_all_bit() {
    int i, res;
    for(i = 0; i <= NUM/32; i++) {
        if (a[i]){
            return a[i];
        }
    }
    return a[i];
}
int list_bit(int n) {
    int i = 0, j = 0;
    for (i = 0; i <= NUM - 1; i++) {
        if (testbit(i)){
            if (j == n) {
                return i;
            }
            j++;
        }
    }
    return i;
}

/* IMPLEMENTATION */
int main() {
    /* create file data.db to save number */
    int fd1 = open("data.db", O_RDWR | O_CREAT | O_TRUNC, 0664);
    if (-1 == fd1) {
        perror("open");
        return -1;
    }

    /* read number of the file, and set flags */
    int i, res;
    for (i = 0; i <= NUM - 1; i++) {
        setbit(i); /* set flags */
    }

    srand(time(0));
    int n = NUM;
    int node;
    printf("starting to generate datas...\n");
    while (test_all_bit()) {
        if (n < 1000 ){
            node = rand() % n;
            i = list_bit(node);
        } else {
            i = rand() % NUM;
        }
        if (testbit(i)) {
            write(fd1, &i, sizeof(int));
            printf("%d\r", i);
            cl_bit(i);
            n--;
        }
    }
    printf("generated ok!\n");

    close(fd1);
    return 0;
}

/*
 * sort.c
 */
/* Include Files */
#include <stdio.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <errno.h>
#include <unistd.h>
#include <time.h>

/* MACROS */
#define    NUM        10000000
#define    SHIFT    5
#define MASK    0x1F

/* CONSTANTS AND VARIABLES */
int a[1 + NUM/32] = { 0 }; /* define an arry and initialize every element */

/* INTERNAL FUNCTIONS */
void setbit(int n) {
    /* find the array member of bit n, and set flag */
    if (n > NUM){
        printf("too large\n");
        return ;
    }
    a[n >> SHIFT] |= 1 << (n & MASK); /* a[n/32] |= 1<<(n%32) */
}
int testbit(int n) {
    /* if the n bit exist, resturn 1 */
    return a[n >> SHIFT] & (1 << ( n & MASK));
}

void show_time() {
    time_t t = time(NULL);
    struct tm *mytm = localtime(&t);
    printf("%04d年%02d月%02d日 %02d:%02d:%02d t=%d\t",
                mytm->tm_year+1900,    mytm->tm_mon+1, mytm->tm_mday,
                    mytm->tm_hour, mytm->tm_min, mytm->tm_sec, (int)t);
}
/* IMPLEMENTATION */
int main() {
    /* open data.db and create file res.db to save the result of sorting */
    int fd1 = open("data.db", O_RDONLY);
    if (-1 == fd1) {
        perror("open");
        return -1;
    }

    /* read number of the file, and set flags */
    int i, n, res;
    show_time();
    printf("start to sorting %d nums...\n", NUM);
    for (i = 0; i <= NUM - 1; i++) {
        res = read(fd1, &n, sizeof(int));
        if (res == -1) {
            perror("read");
            return -1;
        } else if (res == 0) {
            break;
        }
        setbit(n); /* set flags */
    }

    /* traversal the array and write every element to res.db */
    int fd2 = open("res.db", O_RDWR | O_CREAT | O_TRUNC, 0664);
    if (-1 == fd2) {
        perror("open");
        return -1;
    }

    printf("writing to file...\n");
    for (i = 0; i <= NUM -1; i++) {
        if (testbit(i)) {
            write(fd2, &i, sizeof(int));
            //printf("%d\t", i);
        }
    }
    show_time();
    printf("sorting over!\n");

    close(fd1);
    close(fd2);
    return 0;
}

gcc generate.c -o generate
gcc sort.c -o sort
然后分别运行generate可以生成10000000个数字,sort可以完成排序

算法是没有固定的,就好比是说飞机快还是汽车快;
算法要靠平时慢慢的积累;