How to read a binary file in C

Hi Sudhakar,

FILE *stream;
   char buffer[513];
   int  i, ch;

   /* Open file to read line from: */
   if( (stream = fopen( "c:\\a.doc", "rb" )) == NULL )
      exit( 0 );

   /* Read in first 80 characters and place them in "buffer": */     <<<<< 512 bytes
   ch = fgetc( stream );
   for( i=0; (i < 513 ) ; i++ )
   {
      buffer[i] = (char)ch;
      ch = fgetc( stream );
       printf("\n%d %d",i,ch);    <<<<<< you will get equivalent int printed here ... good enough ... Note that first character read will not be printed.
   }

   /* Add null to end string */
   buffer[i] = '\0';       <<<<<<<<<<<, not a good idea ... 0 and other such characters and character sequences can occur anywhere within binary data ... Depending on a terminator character for binary data is definitely bad idea ... You will be better off with the sound knowledge of the number of chracters you have read

   printf( "%s\n", buffer );    >>>>> Since the file is binary, most likely you will get gibbersih which will terminate as soon as 0 occurs in the buffer ... That is the way strings (%s) behaves .... remember that this is not a string but binary data ... If you wish tp print it, use fwrite instead ... Something like
fwrite ( buffer, 1, 513, stdout);
stdout, is defined in stdio.h

   fclose( stream );

cheers
sunnycoder

Hello All,

Thanks for the immediate responses, but the problem is still continuing

//Reading methods
int IOUtils::readFully(FILE* in,unsigned char b[],int length )
{
      return readFully(in,b,0,length);
      return 0;
}

int  IOUtils::readFully(FILE* in, unsigned char b[] , int off, int len)
{
                    
            //int got=fread(b,sizeof(unsigned char), 1,in);
             int got=fread(b, 1,sizeof(b),in);
             return got;
                  

}



//implemetation methods

HeaderBlockReader::HeaderBlockReader(char* file)
{
      FILE* fp=fopen(file,"rb");
      //int fp=_open(file,O_RDONLY);
      if(fp>0)
      {
            unsigned char* _data=new unsigned char[512];
            _data = new unsigned char[512];
          IOUtils *utils=new IOUtils();
        int i = utils->readFully(fp, _data,512);
            printf("%d",i);
            if(i!=512)
            {
                  printf(" nota microsoft document");
            }
            else
            {
                  LongField *longfield = new LongField(0, _data);
                  
                  printf(" Possible it could be a Microsoft Document %x",longfield->get());
            }
      }
      else
      {
                        printf(" Document does not exists");
      }
      fcloseall();
}

this is C++ ... I would suggest sticking to either C or C++ ... you are mixing the C++ code with C programming style

Yes I do agrre,

When I worked on this previously, i never faced the problems.

Now this is immediate requirement for me. to integrate c,c++ and java files and strucking in file reading concepts :(

check http:#11322790  for reason why your code was not working and http:#11322746 for a working example

>>   int got=fread(b, 1,sizeof(b),in);

but i am doing the same know? Then what is the reason for not working

>Then what is the reason for not working

What is the output you see ?

instead of reading all 512 bytes it is just reading 6 bytes

As you said, intgrating c and c++ may gives problems

I just done a test run on document of size 19kb. but the out put i got is of 20 kb

here is the test program

#include <io.h>
#include <stdio.h>

void main()
{
       FILE *in = fopen("c:\\a.doc","rb");
      FILE *out = fopen("c:\\out_file.doc","w");

      int len;
      char buffer[1024];

      while((len = fread(buffer,1,sizeof(buffer),in)) > 0)
      {
         fwrite(buffer,1,len,out);
      }

      fclose(in);
      fclose(out);
}

Nope, it is just printing 6 bytes ... see my comment http:#11322790

   /* Add null to end string */
   buffer[i] = '\0';       <<<<<<<<<<<, not a good idea ... 0 and other such characters and character sequences can occur anywhere within binary data ... Depending on a terminator character for binary data is definitely bad idea ... You will be better off with the sound knowledge of the number of chracters you have read

   printf( "%s\n", buffer );    >>>>> Since the file is binary, most likely you will get gibbersih which will terminate as soon as 0 occurs in the buffer ... That is the way strings (%s) behaves .... remember that this is not a string but binary data ... If you wish tp print it, use fwrite instead ... Something like
fwrite ( buffer, 1, 513, stdout);
stdout, is defined in stdio.h

>I just done a test run on document of size 19kb. but the out put i got is of 20 kb
try comparing the files ... Are you sure you are not seeing size on disk for doc2 ... Also, which platform and which compiler ?

>>buffer[i] = '\0';       <<<<<<<<<<<, not a good idea ...

after removing that code only i am getting this error

I am compiling the program on visual studio (VC++ 6.0)

when i try to open the output document with ms word, ms word shows an error alert saying the document is corrupted

Instead of following the below
while((len = fread(buffer,1,sizeof(buffer),in)) > 0)

I use the below and worked fine.
while((_len = fread(buffer,1,len,in)) > 0)

Any specific reason for this. In both the cases he initial size i allocated is 512 only.

thanks
sudhakar

>i allocated is 512 only.

Did you malloc it ?

@sunny
as my program is combination of c++ and c i used new operator for this

@ankuratvb

Yes, that's true

sizeof (char *) is 4 ... sizeof (char [] ) is actual size of the array

Never mind ... using len instead of sizeof is safer and less confusing ... so stick to it

Also,

in your fwrite(),the 2nd and 3rd arguments should be interchanged.

See:
http://bama.ua.edu/cgi-bin/man-cgi?fwrite+3C

the 2nd arg. is the size of the buffer and the 3rd arg. is the number of items to be written.

That's it,

I understand the problem. There is a lot's of difference in new and malloc. I just forget that.

That is the reason you said combining both languages in not a good idea.

thanks
sudhakar

Ankur,

>the 2nd arg. is the size of the buffer and the 3rd arg. is the number of items to be written.

from the man page

 size_t   fwrite(const  void  *ptr,  size_t  size,  size_t  nmemb,  FILE
       *stream);

DESCRIPTION
       The function fread reads nmemb elements of data, **each size bytes  long**,
       from  the  stream  pointed  to  by stream, storing them at the location
       given by ptr.

Yeah.My Mistake.

I thought that you were trying to read the entire buffer at one go,but strangely that doesnt work.

If you read one byte at a time,it works.

sunny

      char* str=new char[100];
      char str1[100];
      char *str2=(char*)malloc(100);

      printf("\n\n\n\%d   %d  %d", sizeof(str), sizeof(str1), sizeof(str2));


I just tested the above.  Compiler should return value 100 or 4 for this?

>     printf("\n\n\n\%d   %d  %d", sizeof(str), sizeof(str1), sizeof(str2));
Its been too long since I worked with C++ so not sure for str, but I guess sizeof(str) should be 4

str1 should be 100 and str2 should be 4

Thanks to all of you.

You helped me alot. ow my coding will move fastly. :-)

Hi,

What i was trying to get through was that instead of:
>fwrite(buffer,1,len,out);

You can use:

fwrite(buffer,len,1,out);

This way you'll be writig len number of bytes in one go,while in the former case,you'll be writing one byte at a time,repeating this len times.

I tested it out and it works.

>This way you'll be writig len number of bytes in one go,while in the former case,you'll be writing one byte at a time,repeating
>this len times.
Memory locations in either case are recontiguous and are handled in the same way ... You wont be performing mulitple writes with the former

It says, write n items of size 1 byte starting from location x and not write 1 byte n times starting from location x ;o)

As always,sunny,you are right. :~)

I took the pains of timing the two programs and they take about the same time.

:)

Hello you can try like this

int main () {
  FILE * pFile;
  long lSize;
  char * buffer;

  pFile = fopen ( "myfile.txt" , "rb" );
  if (pFile==NULL) exit (1);

  // obtain file size.
  fseek (pFile , 0 , SEEK_END);
  lSize = ftell (pFile);
  rewind (pFile);

  // allocate memory to contain the whole file.
  buffer = (char*) malloc (lSize);
  if (buffer == NULL) exit (2);

  // copy the file into the buffer.
  fread (buffer,1,lSize,pFile);

  /*** the whole file is loaded in the buffer. ***/

  // terminate
  fclose (pFile);
  free (buffer);
  return 0;
}