A Strategy for Displaying Numeric Data with Significant Figures

Introduction

Problem: Some times numeric data, for example floats and doubles, needs to be displayed with a certain amount of decimal precision and at times with a given amount of significant figures.

For example:

4.5568 displayed with 2 Significant Figures is: 4.6
0.0221 displayed with 3 Significant Figures is: 0.022
5.9 displayed with 1 Significant Figures is: 6

Strategy: Using the log10 function we can calculate the number of digits for the integer part of the number we are manipulating. The decimal part of the number can be displayed properly with a sprintf call or something similar as we can see in our function below. This solution also rounds the final number, which is usually desirable when displaying significant figures.

double CalculateSigFig(double num, int sigfig)
{
    char buff[128];//temp buffer
    double temp = 0; 
    int lognum = 0;
    std::string formatbuff;//format string

    if(sigfig <= 0 || num == 0)//damage control
        return num;

    temp = int(num);
    if(temp * -1 > 0)//check for negative numbers
        temp *= -1;

    lognum = (int)log10(temp);//get significant digits of integer part of num
    ++lognum; //adjust for log10 results
    sigfig -= lognum;//subtract for decimal precision calculation

    if(sigfig < 0)//check for negative results
        sigfig = 0;


    //construct format string
    formatbuff = "%";
    itoa(lognum, buff, 10);//convert lognum to int
    formatbuff += buff;
    formatbuff += ".";
    itoa(sigfig, buff, 10);
    formatbuff += buff;
    formatbuff += "f";
    //end construct format string

    //use format string to get desired results
    sprintf(buff, formatbuff.c_str(), num);

    return atof(buff);//return double converted from string
}

This article has no explicit license attached to it but may contain usage terms in the article text or the download files themselves. If in doubt please contact the author via the discussion board below.

A list of licenses authors might use can be found here