DaniWeb Forum Index > Software Development > C++

C fputc and fgetc vs C++ read and write

Please support our C++ advertiser: Intel Parallel Studio Home

•

Join Date: Aug 2005

Posts: 148

Reputation: Micko is on a distinguished road

Solved Threads: 6

Micko

Offline

Junior Poster

C fputc and fgetc vs C++ read and write

Aug 31st, 2005

Hello,
copying from one file to another is a simple matter.
In C it's something like this:

 Help with Code Tags 
 C++ Syntax (Toggle Plain Text) 
 
int main(void)
{
	FILE * fp1, * fp2 ;
	int buff;
 
	fp1 = fopen ("name", "rb");
	fp2 = fopen ("name", "wb");
 
	while ( (buff = fgetc(fp1)) != EOF)
	{
		fputc(buff, fp2);
	}
 
}
int main(void)
{
	FILE * fp1, * fp2 ;
	int buff;

	fp1 = fopen ("name", "rb");
	fp2 = fopen ("name", "wb");

	while ( (buff = fgetc(fp1)) != EOF)
	{
		fputc(buff, fp2);
	}

}

I'm wondering how this is done in C++
My idea is this:

 Help with Code Tags 
 C++ Syntax (Toggle Plain Text) 
int main()
{
	int buf;
 
	ifstream fp_in("my.txt", ios_base::binary | ios_base::in );
	ofstream fp_out("my_copy.txt",	ios_base::binary | ios_base::out);
 
	while (fp_in.read ( (char*) &buf, sizeof buf ))
		fp_out.write ( (char*) &buf, sizeof buf );
}
int main()
{
	int buf;

	ifstream fp_in("my.txt", ios_base::binary | ios_base::in );
	ofstream fp_out("my_copy.txt",	ios_base::binary | ios_base::out);

	while (fp_in.read ( (char*) &buf, sizeof buf ))
		fp_out.write ( (char*) &buf, sizeof buf );
}

I think this is very bad solution because of performance.
Can you suggest me any other way in C++ to achive this?

Thanks

•

Join Date: Aug 2005

Posts: 15,648

Reputation: Ancient Dragon has a reputation beyond repute

Solved Threads: 1498

Ancient Dragon Ancient Dragon is online now

Online

Still Learning

Re: C fputc and fgetc vs C++ read and write

Aug 31st, 2005

only write out the number of bytes read. At some point (at end-of-file) the number of bytes read will be less than the buffer size. This would go much faster if you read/write with larger buffer size -- say 255 characters.

Help with Code Tags

(Toggle Plain Text)

int main()
{
	char buf[255];

	ifstream fp_in("my.txt", ios_base::binary | ios_base::in );
	ofstream fp_out("my_copy.txt",	ios_base::binary | ios_base::out);

	while (fp_in.read ( buf, sizeof buf ))
		fp_out.write (buf, fp_in.gcount() );
}

•

Join Date: Sep 2004

Posts: 7,850

Reputation: Narue has a reputation beyond repute

Solved Threads: 754

Narue

Offline

Senior Bitch

Re: C fputc and fgetc vs C++ read and write

Aug 31st, 2005

>copying from one file to another is a simple matter.
On the surface at least.

>I'm wondering how this is done in C++
A direct translation of C streams to C++ streams is:

 Help with Code Tags 
 C++ Syntax (Toggle Plain Text) 
#include <fstream>
 
int main()
{
  ifstream in ( "infile" );
  ofstream out ( "outfile" );
 
  if ( in && out ) {
    char ch;
 
    while ( in.get ( ch ) )
      out.put ( ch );
  }
}
#include <fstream>

int main()
{
  ifstream in ( "infile" );
  ofstream out ( "outfile" );

  if ( in && out ) {
    char ch;

    while ( in.get ( ch ) )
      out.put ( ch );
  }
}

>I think this is very bad solution because of performance.
No, it's a very bad solution because it's a very bad solution.

You use an int when you should be using a char, to be consistent with the C program, or an array of char, to do block copies. You'll find that people think a character by character copy (such as the one above) is expensive because they think that each character is read and written individually, thus causing extra device operations. In reality, the stream will almost always buffer the data in the most efficient way possible to avoid device reads just as you would if you copied blocks manually. Only complain about performance if you can prove that there's actually a bottleneck.

New members chased away this month: 4

•

Join Date: Aug 2005

Posts: 148

Reputation: Micko is on a distinguished road

Solved Threads: 6

Micko

Offline

Junior Poster

Re: C fputc and fgetc vs C++ read and write

Aug 31st, 2005

•

Originally Posted by Narue

>copying from one file to another is a simple matter.
On the surface at least.

>I'm wondering how this is done in C++
A direct translation of C streams to C++ streams is:
Help with Code Tags

C++ Syntax (Toggle Plain Text)
#include <fstream>
 
int main()
{
  ifstream in ( "infile" );
  ofstream out ( "outfile" );
 
  if ( in && out ) {
    char ch;
 
    while ( in.get ( ch ) )
      out.put ( ch );
  }
}
#include <fstream> int main() { ifstream in ( "infile" ); ofstream out ( "outfile" ); if ( in && out ) { char ch; while ( in.get ( ch ) ) out.put ( ch ); } }
>I think this is very bad solution because of performance.
No, it's a very bad solution because it's a very bad solution. You use an int when you should be using a char, to be consistent with the C program, or an array of char, to do block copies. You'll find that people think a character by character copy (such as the one above) is expensive because they think that each character is read and written individually, thus causing extra device operations. In reality, the stream will almost always buffer the data in the most efficient way possible to avoid device reads just as you would if you copied blocks manually. Only complain about performance if you can prove that there's actually a bottleneck.

Great, thanks, I was thinking about get and put, but I wasn't sure about text and binary mode. In C implementation with fputc and fgetc I opened files in binary mode, so I wanted to use binary mode also in C++ implementation... Or maybe that's not matter here?