Tip: How to avoid string parsing, File I/O

[Clinton Portis]

A lot of people have questions about how to load a text file using file I/O. There seems to be 2 popular methods to loading a text file:

Use fstream's >> extraction operator. Simple enough, load the file directly into individual containers, word at a time, but you lose your delimiting white spaces (assuming you need them). So some file integrity is lost, unless you make the effort re-construct the file by re-inserting all the white spaces.

So instead, you decide to use getline(). You preserve all white spaces and load the file line by line... but now you have to parse a line of data into individual substrings; either by using <string> member functions, performing string[] array operations, or using strtok().

One alternative I would like to suggest: why not do both? It is possible to read a file in it's entirety and read in text 'word at a time' into individual containers.. without having to do any string parsing:

#include<string>

//Be sure to open your file in binary mode
infile.open("C:\\Users\\Dave\\Documents\\test.txt", ifstream::binary);

//Here you can load the file in it's entirety    
while(getline(infile, lines[i]))
{
     //Go back to the start of the line
     infile.seekg(begin, ios::beg);

     //Now you can load the same data into individual containers
     for(int j=0; j<4; j++)
     {
          infile >> words[word_count];
          word_count++;
     }

     //Discard any extra characters left behind
     infile.ignore(100, '\n');

     //Save current position, so you can go back to the beginning of next line
     begin = infile.tellg();

     i++;
 }

So now you have the entire document preserved in lines[], and you have individual line contents stored in words[]. Depending on your application needs, this method might be a viable option in that you get the best of both worlds without having to do any string parsing.

[mrnutty]

Don't forget, that there are also other methods. For example, you can
extract character by character.

Here is an example of that :

#include<iostream>
#include<string>
#include<fstream>

using namespace std;


int main()
{
	ifstream iFile("test.txt");
	if(!iFile) return -1;

	char ch = 0;

	string content = "";

	while(iFile.get(ch) ){
		content += ch;		
	}

	cout << content << endl;

	return 0;
}

And if you want only the words in the file, then a simple if statement
to check if the char is a space and also using vectors of string will do.

[Ancient Dragon]

you could have just used stringstream to split the line into words without re-reading the file. And it would have been more useful to use vector to hold the lines/words.

[edit]I don't see any reason at all for reading the file one character at a time. Too much work just to extract individual words.

[mrnutty]

[edit]I don't see any reason at all for reading the file one character at a time. Too much work just to extract individual words.

Yea I know. Just throwing the possibility out there. Reading
character by character might be helpful in some cases though, like
frequency counter, or whatever.