搬：超过6G的大文件是如何读取的,附源代码

超过6G的大文件是如何读取的,附源代码。含有网上很少有论述的信息

2017年12月29日 10:47:28 lzyzuixin 阅读数：1144

https://blog.youkuaiyun.com/lzyzuixin/article/details/78928915

 

因为工作关系，需要从大文件中筛选数据，进行比对。文件大小一般都在6g左右。

读取大文件有如下两种方法，一是用fopen打开文件，fgetline循环读取，fclose关闭文件；二是用open打开函数，用lseek获取文件大小,用mmap大文件内存映射，用munmap关闭内存映射，用close关闭文件句柄。方式一教慢，就不再详细描述。主要描述方式二。

方式二，网上介绍也有很多，但是鲜有介绍当大于4G后，读取方法。用long型读取文件大小时，最多是4294967295个字节，也就是4G。解决方法是用long long来读取文件的大小。

 

样例代码如下：

#include<stdio.h>
#include<string.h>
#include<stdlib.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/time.h> 
#include <sys/mman.h> 
#include <sys/types.h>
#include <errno.h>
#include <unistd.h>
using namespace std;

#include "dlist.h"

//呼叫流程数组
typedef struct s_callflow
{
   char identifier[100];
   char billingid[200];
   bool response;
   s_callflow()
   {
      strcpy(identifier,"");
      strcpy(billingid,"");
      response=false;
   }
} callflow;


callflow *call_instance(char *identifier,char *billingid ,bool response) 
{ 
    callflow *call_ptr; 
    call_ptr = (callflow *)malloc(sizeof(callflow)); 
    if( call_ptr==NULL ) 
        return NULL; 
   
    strcpy(call_ptr->identifier,identifier);
    strcpy(call_ptr->billingid,billingid);
    call_ptr->response = response;
    return call_ptr; 
} 

void updatecall(DList *list,char *identifier)
{
    if(list==NULL) return;
    DListElmt *new_element;
    new_element = list->tail;   
    while(new_element!=NULL)
    {
        callflow * flow = (callflow *)new_element->data;
        if(flow!=NULL)
        {
            if(strcmp(flow->identifier,identifier)==0 && flow->response==false)
            {
                flow->response=true;
                break;
            }
        }
        if( new_element == list->head )
        {
            break;
        }
        new_element = new_element->prev;
    }
  
}

/*destroy */ 
void destroy(void *data) 
{ 
    free(data); 
    return; 
} 

void output(DList *list,bool calling)
{
    if(list==NULL) return;
    DListElmt *new_element;
    new_element=list->head;
    int count = list->size;
    int response=0;
    FILE *out;
    if(calling==1)
    {
        out=fopen("calling.txt","w");
    }
    else
    {
        out=fopen("called.txt","w");
    }
    char buffer[255];
    while(new_element!=NULL)
    {
         callflow * flow = (callflow *)new_element->data;
         if(flow!=NULL)
         {
            if(flow->response==true)
            {
               response++;
            }
            else
            {
                 sprintf(buffer,"billingid=%s identifier=%s\n",flow->billingid,flow->identifier);
                 fwrite(buffer,strlen(buffer),1,out);
            }
         }
         new_element=new_element->next;
    }
     sprintf(buffer,"count=%d response=%d\n",count,response);
     fwrite(buffer,strlen(buffer),1,out);
     fclose(out);
}


//
int main(int argc,char * argv[])
{
   //size_t lsize=0;
   long long lsize=0;
   const char *localpc="16592304";
   DList calling_node;
   DList called_node;
   dlist_init(&calling_node, destroy);
   dlist_init(&called_node, destroy); 
   char opc[3][200];
   char dpc[3][200];
   char identifier[2][100];
   char billingID[2][200];
   //FILE * fp=NULL;
   //fp = fopen(argv[1],"r");
   char *pBuffer=NULL; 
   char *pStart=NULL,*pEnd=NULL; 
   int fd = open(argv[1],O_RDONLY);
   //size_t nFileSize=0; 
   //size_t nOffset=0; 
   //size_t nLineAmount=0; 
   //struct stat fileState; 
   //fstat(fd,&fileState); 
   //nFileSize=fileState.st_size;
   long long nFileSize=0;
   long long nOffset;
   long long nLineAmount;
   nFileSize =(long long)lseek(fd,0,SEEK_END);
   //nFileSize = (unsigned long)fileState.st_size;
   pBuffer=(char *)mmap(NULL,nFileSize,PROT_READ,MAP_SHARED,fd,0); 
   pEnd=pStart=pBuffer;
   char line[2048];
   int flag;
   flag=-1;
   int in=-1;
   printf("nFileSize=%lld\n",nFileSize);
   int load = 0;
   int preload = 0;
   while(lsize<nFileSize-4)
   {
       lsize++;
       load = (int) (((float)lsize / (float)nFileSize) * 100);
       if(preload!=load)
       {
          printf("%3d \n",load);
          preload = load;
          sleep(1);
          printf("\b");
       }
       char a = *pEnd;
       if(a==13)
       {
          char b = *(++pEnd);
          if(b==10)
          {
               int len = pEnd-pStart;
               if(len<1) break;
               if(len>2047) break;
               strncpy(line,pStart,len);
               pStart=pEnd+1;
               if(strstr(line," OPC")!=NULL)
               {
                   sscanf(line,"%s%s%s",opc[0],opc[1],opc[2]);
               }
               else if(strstr(line," DPC")!=NULL)
               {
                   sscanf(line,"%s%s%s",dpc[0],dpc[1],dpc[2]);
               }
               else if(strstr(line,"queryWithPerm")!=NULL)
               {
                  flag=0; //
               }
               else if(strstr(line,"response")!=NULL)
               {
                  flag=1; //
               }
               else if(strstr(line,"identifier")!=NULL)
               {
                   sscanf(line,"%s%s",identifier[0],identifier[1]);
                   //
                   if(strcmp(opc[1],localpc)==0)
                   {
                       //
                       if(flag==1) //response
                       {
                            //找主叫流程
                            updatecall(&calling_node,identifier[1]);
                            //找被叫流程
                            updatecall(&called_node,identifier[1]);
                            flag=-1;
                       }
                   }
               }
               else if(strstr(line,"originationRequest")!=NULL)
               {
                   //主叫流程
                   if(strstr(line,"originationRequestRes")==NULL)
                   {
                       in = 1;
                   }
                  
               }
               else if(strstr(line,"billingID")!=NULL)
               {
                  sscanf(line,"%s%s",billingID[0],billingID[1]);
                  if(in==1)
                  {
                     callflow * calling = call_instance(identifier[1],billingID[1] ,false);
                     dlist_ins_next(&calling_node,calling_node.tail,(void *)calling);
                     in = -1;
                  }
               }
               else if(strstr(line,"initial-Termination (38)")!=NULL)
               {
                   //被叫流程
                   callflow * called = call_instance(identifier[1],billingID[1] ,false);
                   dlist_ins_next(&called_node,called_node.tail,(void *)called);
               }
          }
       }
       else
       {
          pEnd++;
       }
     
   }
   output(&calling_node,true);
   output(&called_node,false);
   dlist_destroy(&calling_node);
   dlist_destroy(&called_node);
   munmap(pBuffer,nFileSize);
   //fclose(fp);
   close(fd);
}