Thread: ifstream text file search for word and gather block of data

Note that when you getline(), you get the whole line of text. When you infile >> imageNumber;, it gets the number after the previous line of text. The number you want is in lineOfText.

Code:

#include <iostream>
#include <string>

int main()
{
   std::string searchString("image");
   std::string lineOfText("image 0");
   std::string::size_type i = lineOfText.find(searchString, 0);
   if (i != std::string::npos)
   {
      std::cout << lineOfText.substr(i + searchString.size() + 1) << std::endl;
   }

   return 0;
}

Originally Posted by Geek10

I never like coming to people for help with empty hands

This is a good attribute to have when asking questions here. People will appreciate it.

Hi pianorain,

Thanks for your input! I have muddled through your code for the past hour and finally realised that:

Code:

std::string::size_type i = lineOfText.find(searchString, 0);

means
1. Starting from the very beginning of the string (0),
2. find the string searchString in the string lineOfText
3. store the location of that string to size_type i.

So I have finally worked out what your code fully means, I think. Below is it expanded, and commented. Hopefully others find it useful as well.

Code:

#include <iostream>
#include <string>
#include <conio.h>
#include <sstream>
//#include <math.h>
//#include <stdio.h>


int main()
{
   std::string searchString("image");
   std::string lineOfText("image 236");
	
	// size_type i is the location of the first string matching searchString found by .find
	std::string::size_type i = lineOfText.find(searchString, 0);
   // searchString.size() is the number of letters in searchString = 5
	std::cout << "searchString.size() " << searchString.size() << "\n";
   // string::npos signals the end of the string?
	std::cout << "std::string::npos " << std::string::npos << "\n";
   std::cout << "i " << i << "\n";
	if (i != std::string::npos) 
   {	// If the location of the first string is not at the end of the string
      std::cout << "The string after the word image: " << lineOfText.substr(i + searchString.size() + 1) << std::endl;
   }
   
	// This makes a string out of the characters following the space after "image"
	std::string numberString = lineOfText.substr(i + searchString.size() + 1);
	// stringstream provides an interface to manipulate strings as if they were input/output streams.
	std::stringstream numberStream(numberString);
	
	
	int imageNumber;	// Initialise the integer for the found value
	if((numberStream >> imageNumber).fail()){	// Convert stream to integer
 	    std::cout << "ERROR: Failed string to integer conversion";
   }
  	std::cout << "imageNumber integer: " << imageNumber << std::endl;

	getch();
   return 0;
}

Thanks so much pianorain, this thread can be marked solved (if this forum does that).

Sorry about the wonky code there - it looks horrible - my code editor (Dev C++) is playing up a bit - I must have a setting wrong somewhere.
Regards,

Geek10

Originally Posted by msh

Code:

// It's better to do this...
while (1) {
    ...
    if (condition) break;
    ...
}

// ...rather then this.
for (;;) {
    ...
    if (condition) break;
    ...
}

This is a matter of subjective style, and I disagree that the former is better than the latter.

Originally Posted by msh

Code:

// Or, in your case...
while (!inFile.eof()) {
    ...
}

No, that is wrong. Unless you take care to do it properly, eof() should not be used to control a loop. Rather:

Code:

while (getline(inFile, lineOfText)) {
    // The eof() check is no longer needed.
    // ...
}

Originally Posted by msh

Is this entirely subjective, or is there a reason for using for( ;; )? I've seen it used, and I understand that it's valid, but I couldn't understand why it's used over while(1).

A long time ago in a galaxy far, far away... the for loop version was better than the while loop version because compilers might evaluate the constant expression instead of recognising the infinite loop.

Just having a look at this topic for reference about searching files through fstream, the code looks quite nice here, and I now understand the concept.

Thanks!

Also, I noticed that the full code was never posted quite completely, so I have taken the liberty to do so below.

NOTE: I have also removed an unneeded library and fixed up the indentation;

Code:

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

using namespace std;

int main(){
    int imageNumber = -1;

    ifstream inFile("data.txt");
    string searchString = "image";
    string lineOfText = "image 236";

    // size_type i is the location of the first string matching searchString found by .find
    string::size_type i = lineOfText.find(searchString, 0);

    // searchString.size() is the number of letters in searchString = 5
    cout << "searchString.size() " << searchString.size() << "\n";

    // string::npos signals the end of the string?
    cout << "string::npos " << string::npos << "\n";
    cout << "i " << i << "\n";

    if (i != string::npos){    // If the location of the first string is not at the end of the string
        cout << "The string after the word image: " << lineOfText.substr(i + searchString.size() + 1) << endl;
    }

    // This makes a string out of the characters following the space after "image"
    string numberString = lineOfText.substr(i + searchString.size() + 1);

    // stringstream provides an interface to manipulate strings as if they were input/output streams.
    stringstream numberStream(numberString);


    imageNumber=0;    // Initialise the integer for the found value
    if((numberStream >> imageNumber).fail()){    // Convert stream to integer
        cout << "ERROR: Failed string to integer conversion";
    }
    cout << "imageNumber integer: " << imageNumber << endl;

    cout << endl << "Press ENTER to continue" << endl;
    cin.clear();
    cin.sync();
    cin.ignore();
}

Originally Posted by Geek10

What does your cin.sync do? According to C++ Reference, "Synchronizes the buffer associated with the stream to its controlled input sequence. This effectively means that the unread characters in the buffer are discarded."

Why are characters unread, let alone discarded?

User input is stored in a buffer to be read, so the idea here is to discard whatever is left on the input buffer, and then read and ignore a character, in an attempt (that is generally misguided) to keep the console window open after the process (program) has terminated.

That said, cin.sync() is not guaranteed to do this, i.e., whatever C++ reference you are using, it is giving information that may be usually true, but not guaranteed to be true, hence it is actually not portable.