Read Numeric Data from a Text File in C++

It can depend, especially on whether your file will have the same number of items on each row or not. If it will, then you probably want a 2D matrix class of some sort, usually something like this:

class array2D { 
    std::vector<double> data;
    size_t columns;
public:
    array2D(size_t x, size_t y) : columns(x), data(x*y) {}

    double &operator(size_t x, size_t y) {
       return data[y*columns+x];
    }
};

Note that as it's written, this assumes you know the size you'll need up-front. That can be avoided, but the code gets a little larger and more complex.

In any case, to read the numbers and maintain the original structure, you'd typically read a line at a time into a string, then use a stringstream to read numbers from the line. This lets you store the data from each line into a separate row in your array.

If you don't know the size ahead of time or (especially) if different rows might not all contain the same number of numbers:

11 12 13
23 34 56 78

You might want to use a std::vector<std::vector<double> > instead. This does impose some overhead, but if different rows may have different sizes, it's an easy way to do the job.

std::vector<std::vector<double> > numbers;

std::string temp;

while (std::getline(infile, temp)) {
    std::istringstream buffer(temp);
    std::vector<double> line((std::istream_iterator<double>(buffer)),
                             std::istream_iterator<double>());

    numbers.push_back(line);
}

...or, with a modern (C++11) compiler, you can use brackets for line's initialization:

    std::vector<double> line{std::istream_iterator<double>(buffer),
                             std::istream_iterator<double>()};

The input operator for number skips leading whitespace, so you can just read the number in a loop:

while (myfile >> a)
{
    // ...
}

Repeat >> reads in loop.

#include <iostream>
#include <fstream>
int main(int argc, char * argv[])
{
    std::fstream myfile("D:\\data.txt", std::ios_base::in);

    float a;
    while (myfile >> a)
    {
        printf("%f ", a);
    }

    getchar();

    return 0;
}

Result:

45.779999 67.900002 87.000000 34.889999 346.000000 0.980000

If you know exactly, how many elements there are in a file, you can chain >> operator:

int main(int argc, char * argv[])
{
    std::fstream myfile("D:\\data.txt", std::ios_base::in);

    float a, b, c, d, e, f;

    myfile >> a >> b >> c >> d >> e >> f;

    printf("%f\t%f\t%f\t%f\t%f\t%f\n", a, b, c, d, e, f);

    getchar();

    return 0;
}

Edit: In response to your comments in main question.

You have two options.

  • You can run previous code in a loop (or two loops) and throw away a defined number of values - for example, if you need the value at point (97, 60), you have to skip 5996 (= 60 * 100 + 96) values and use the last one. This will work if you're interested only in specified value.
  • You can load the data into an array - as Jerry Coffin sugested. He already gave you quite nice class, which will solve the problem. Alternatively, you can use simple array to store the data.

Edit: How to skip values in file

To choose the 1234th value, use the following code:

int skipped = 1233;
for (int i = 0; i < skipped; i++)
{
    float tmp;
    myfile >> tmp;
}
myfile >> value;

Tags:

C++

Recent Posts