Converting string with decimal to double using atof in C++

Question

In my program im trying to read in a file with money like this "$876,725.38" I was able to remove the comma and dollar sign and I am now trying to convert the string to a double using atof but the issue im having now is after printing MONEYS after the conversion I get everything before the period on majority of these, but some with less than 6 digits before the period do print either one or two digits after the period. If anyone has a solution to this it would be greatly appreciated. I'm new to C++ and need help. Thanks!

MAIN ISSUE: Converting from string to double, the data after decimal does not print.

    void readFile()
{
    string line, nuMoney, munney, cents;
    unsigned long dollarSign, period;


    ifstream myfile("MONEY.txt");

    int numLines = countLines("MONEY.txt");
    //cout << "NUMLINES: " << numLines << endl;

    if (!myfile.is_open())
    {
        cout << "ERROR: File could not be opened." << endl;
    }

    for (int i=0; i < numLines-1; i++)
    {
        getline(myfile, line, '\n');
        dollarSign = line.find('$');
        period = line.find('.');

        line.erase (remove(line.begin(), line.end(), ','), line.end()); //remove commas

        munney = line.substr(dollarSign+1);
        //cout << munney << endl;

        //cents = line.substr(period);
        //cout << "Cents: " << cents << endl;

        //double CENTUS = atof(cents.c_str());
        //cout << "CENTUS: " << CENTUS << endl;

        double MONEYS = atof(munney.c_str());
        cout << MONEYS << endl;

        list<double> acct;
        acct.push_back(MONEYS);

    }

}

sample output:

http://stackoverflow.com/help/mcve please!! Inputs, expected outputs and completly compilable code that reproduces the unexpected behavior. — πάντα ῥεῖ, Mar 27 '15 at 19:00
sooo… what's the problem? I don't understand, your description is unclear… — The Paramagnetic Croissant, Mar 27 '15 at 19:02
Storing monetary values as floating points numbers is a really bad idea. Is that a requirement? — stefan, Mar 27 '15 at 19:05
@TheParamagneticCroissant sorry for the bad description... I'm really just having an issue after converting my string to a double. When I print out the double it prints everything before the period but I need it all in order to add it up. Hope that makes some more sense. — Johnny Tran, Mar 27 '15 at 19:06
@JohnnyTran An integer. Just use cents as a unit, not dollars. — stefan, Mar 27 '15 at 19:07
Instead you can use an `int` multiply by 100 to include "cents," divide by 100 to get the monetary value. — Evan Carslake, Mar 27 '15 at 19:07
@stefan what if I was trying to take the whole value like '123456.78'. The `cents` was something I was testing and why its commented out. I really want to keep the whole value together and put it in a STD::LIST after. — Johnny Tran, Mar 27 '15 at 19:15
Take the amount `4.50` as an int so multiply `4.50 * 100 = 450`. Then to convert back `450 / 100 = 4.50` — Evan Carslake, Mar 28 '15 at 00:45

score 1 · Answer 1 · answered Mar 27 '15 at 19:15

Floating point (float, double and long double) aren't great for storing exact values. They are great because they make it easy to do lots of different math on a wide range of values, but there are many values which can't be represented exactly using IEEE floating points. Values as simple as 0.1 can't be exactly expressed by the type, but the input and output routines often round to make it seem like they are handled correctly.

To store decimal perfect values I suggest treating the 2 parts of the value as separate integers when reading in, converting the fractional point based on how many digits after the decimal place were read in, and then doing a little bit of math to convert both pieces into a single integer. For US money the LSB (least significant bit) of this integer may represent 1 cent (unless fractional cents need to be represented accurately). A similar approach can be used for outputting the values.

thank you, im not really sure the math bit to convert the 2 pieces into a single int. Some more insight into that please. — Johnny Tran, Mar 27 '15 at 19:23

score 0 · Answer 2 · answered Mar 27 '15 at 19:21

0

It looks like it's due to default precision of cout stream for floating point values: http://www.cplusplus.com/reference/ios/ios_base/precision/

answered Mar 27 '15 at 19:21

Alexandr Zarubkin

899
1
11
26

score 0 · Accepted Answer · answered Mar 27 '15 at 19:25

0

You have to tell the output stream that you want a fixed precision of 2 digits after the dot:

#include <iomanip>
// ...
std::cout << std::fixed << std::setprecision(2) << MONEYS << '\n';

answered Mar 27 '15 at 19:25

fredoverflow

256,549
94
388
662

Converting string with decimal to double using atof in C++

3 Answers3