Reading white spaces....

[AutoC]

Hi,

Here's my situation.I'm reading a file that consists of characters including white space.Once I read them I'm counting the occurrences of each character.This means I need to count occurrences of white space as well.If I use ifstream I skip the white spaces and I dunno the filesize so I cant use getline.What do I do?

[Faldrax]

You need to use the unformatted input functions, such as read.
You can either read 1 character at a time and count, or read a block into a buffer then count the characters in the buffer (more efficient as less I/O). If doing the latter you will need to use gcount to find out how many characters you did read when you reach end of file.
See http://www.cplusplus.com/reference/iostream/ifstream/ for info on the various functions, etc.

[Mitsakos]

I think you can do it with the getline function fine... The getline stops at the end of the line, only if you use the ">>" operator stops at white space.

I did it like that:

#include <iostream>
#include <fstream>
#include <string>
#include <algorithm>
using namespace std;

int main(){
   fstream inText( "text.txt", ios::in );
   string line, file;
   while( getline( inText, line ) )
      file += line + '\n';
   cout << "White spaces: " << count( file.begin(), file.end(), ' ' );
}

Hope this helps

[EDIT]: Something else I just thought about.
If you don't wont to store all the file in a string (because of ram) then you can use a map ( <char, int> ) and read each letter and store it in there.

[helios]

^Too slow. read() byte by byte and increment the elements of an unsigned[256].

[Mitsakos]

What about the map?

#include <iostream>
#include <fstream>
#include <map>
using namespace std;

int main(){
   fstream inText( "text.txt", ios::in );
   map<char, int> chars;
   char ch;
   //Count letters
   while( ( ch = inText.get() ) && inText.good() )
      ++chars[ch];
   //Display counted letters
   for( map<char, int>::iterator it = chars.begin(); it != chars.end(); ++it )
      if( it->first == '\n' )
         cout << "New lines: " << it->second << " times\n";
      else
         cout << "'" << it->first << "'" << ": " << it->second << " times\n";
}

I did understand what you said but I was wondering which one is better...

[helios]

Maps take O(log n) time to return a reference to an element. Arrays take O(1) time.

Ask this question: "is my map using integral, consecutive keys? Should every key always be present?"
If this is the case, use an array.

[Mitsakos]

that's good to know, thanx