Convert floating point number to string with some formatting

[ssigala]

Convert a floating point value to a string, specifying the number of decimals, the decimal separator and the thousands separator.

Code: Select all

// d: value to convert
// decimals: number of decimals to output (<= 9)
// usecomma: use the comma as decimal separator if true, otherwise the dot
// thousands_spacer: separate the thousands if true
wxString DoubleToString(double d, int decimals, bool usecomma, bool thousands_spacer)
{
	if (decimals > 0) {
		double d_int, d_fract;
		d_fract = modf(d, &d_int);
		wxString s_int, s_fract;
		s_int = wxString::Format("%.1f", d_int);
		s_int.RemoveLast(2);
		if (thousands_spacer)
			s_int = FormatThousands(s_int);
		switch (decimals) {
		case 1: s_fract = wxString::Format("%.1f", d_fract); break;
		case 2: s_fract = wxString::Format("%.2f", d_fract); break;
		case 3: s_fract = wxString::Format("%.3f", d_fract); break;
		case 4: s_fract = wxString::Format("%.4f", d_fract); break;
		case 5: s_fract = wxString::Format("%.5f", d_fract); break;
		case 6: s_fract = wxString::Format("%.6f", d_fract); break;
		case 7: s_fract = wxString::Format("%.7f", d_fract); break;
		case 8: s_fract = wxString::Format("%.8f", d_fract); break;
		case 9: s_fract = wxString::Format("%.9f", d_fract); break;
		}
		s_fract.Remove(0, 2); // remove "0." or "0,"
		return s_int + (usecomma ? ',' : '.') + s_fract;
	} else { // decimals == 0
		double fl = floor(d);
		double ce = ceil(d);
		wxString s;
		if (d - fl < ce - d)
			s = wxString::Format("%d", (int)fl);
		else
			s = wxString::Format("%d", (int)ce);
		if (thousands_spacer)
			return FormatThousands(s);
		return s;
	}
}

Examples:

Code: Select all

wxString s;

s = DoubleToString(3.141592, 4, false, false);
// s now contains "3.1416"
s = DoubleToString(3.141592, 4, true, false);
// s now contains "3,1416"
s = DoubleToString(3141.5926535, 3, false, true);
// s now contains "3,141.593"
s = DoubleToString(3141.5926535, 3, true, true);
// s now contains "3.141,593"

[leio]

You can also simply generate the formatting yourself too instead of an enormouse switch block. Like this (anything else is left as is):

Code: Select all

// d: value to convert
// decimals: number of decimals to output (<= 9)
// usecomma: use the comma as decimal separator if true, otherwise the dot
// thousands_spacer: separate the thousands if true
wxString DoubleToString(double d, int decimals, bool usecomma, bool thousands_spacer)
{
	if (decimals > 0) {
		double d_int, d_fract;
		d_fract = modf(d, &d_int);
		wxString s_int, s_fract;
		s_int = wxString::Format("%.1f", d_int);
		s_int.RemoveLast(2);
		if (thousands_spacer)
			s_int = FormatThousands(s_int);

		// start of modified code
		char *buf[16];
		snprintf(buf, sizeof(buf), "\%.%df", decimals);
		s_fract = wxString::Format(buf, d_fract);
		// end of modified code
		s_fract.Remove(0, 2); // remove "0." or "0,"
		return s_int + (usecomma ? ',' : '.') + s_fract;
	} else { // decimals == 0
		double fl = floor(d);
		double ce = ceil(d);
		wxString s;
		if (d - fl < ce - d)
			s = wxString::Format("%d", (int)fl);
		else
			s = wxString::Format("%d", (int)ce);
		if (thousands_spacer)
			return FormatThousands(s);
		return s;
	}
}

[utelle]

Unfortunately the code does not work properly. Try

Code: Select all

s = DoubleToString(1.99996, 4, false, false);

s now contains "1.0000" , but it should contain "2.0000". Rounding effects have to be taken into account, i.e. d_int has to be incremented by 1 if the first character in s_fract is 1 (it's not always 0 as the code assumes).

Following is some code correctly handling this issue. It can easily be adjusted to use comma or point as the decimal separator and to add thousands separator formatting:

Code: Select all

wxString Double2String(double value, int precision = 0)
{
  wxString number;
  if (precision < 0)
  {
    precision = 0;
  }
  else if (precision > 16)
  {
    // max. precision of type double is about 16 on most architectures
    precision = 16;
  }

  // Use absolute value locally
  double localValue = fabs(value);
  double localFraction = (localValue - floor(localValue)) +(5. * pow(10.0, -precision-1));
  if (localFraction >= 1)
  {
    localValue += 1.0;
    localFraction -= 1.0;
  }
  localFraction *= pow(10.0, precision);

  if (value < 0)
  {
    number += wxString(_T("-"));
  }

  number += wxString::Format(_T("%.0lf"), floor(localValue));

  // generate fraction, padding with zero if necessary.
  if (precision > 0)
  {
    number += wxString(_T("."));
    wxString fraction = wxString::Format(_T("%.0lf"), floor(localFraction));
    if (fraction.Length() < precision)
    {
      number += wxString(_T('0'), precision-fraction.Length());
    }
    number += fraction;
  }

  return number;
}

Regards,

Ulrich

[neo_in_matrix]

Where is FormatThousands? I don't think it's provided by wxWidgets - I could not look it up in the Help.

[utelle]

Where is FormatThousands? I don't think it's provided by wxWidgets - I could not look it up in the Help.

The code was posted in a previous posting of ssigala. Look here:

http://forums.wxwidgets.org/viewtopic.php?t=3578

[JPlaroche]

// Double convert wxString: value to convert
// decimals: number of decimals

wxString DblToStr(double valx , unsigned int decimal=0 )
{

double d_int, d_fract , value;
wxString edt , buf s_fract;
if (decimal > 16 ) decimal = 16;

if (valx< 0 ) value = valx *-1; else value =valx;

d_fract = modf(value , &d_int);

// recuperation decimal
edt =_T("%#." + wxString::Format("%d",decimal)+"f");

s_fract = wxString::Format(edt, d_fract);
//entier +1 si arrondit
if (s_fract[0] == '1') d_int ++;

s_fract.ToDouble(&d_fract);
if (decimal == 0 ) value = d_int ; else value = d_int+ d_fract;
if (valx < 0 ) value *=-1;
buf= wxString::Format(edt,value);
if (decimal == 0 ) buf.RemoveLast(1);
return buf;

}


j'ai test

[lowjoel]

[quote="JPlaroche"]j'ai test

[JPlaroche]

my version is ok DblToStr

lollllllllllll

wxString DoubleToString is not ok

[utelle]

my version is ok DblToStr

No, it isn't. At least not the above version, which not even compiles. After eliminating the syntax error(s) I tested it with

Code: Select all

DblToStr(1.98,1)

giving 3.0.

I wouldn't call that ok, would I?

But the code you posted in http://forums.wxwidgets.org/viewtopic.php?p=27634 seems to be ok.

wxString DoubleToString is not ok

I agree, but the code I posted above, namely the function Double2String (as used in my component wxPdfDocument), works as expected.

Regards,

Ulrich

[JPlaroche]

I have a previous version of pdfdoc
Having updated her I again tested him and it is OK

My excuses. Thank you is for work

with debug I still have some problem
In the recoveries of intermediate values.
I am in version 6.9 wxdevcpp

[utelle]

I have a previous version of pdfdoc
Having updated her I again tested him and it is OK

Well, the code of method Double2String was/is identical in all versions of wxPdfDocument published at wxCode.

with debug I still have some problem
In the recoveries of intermediate values.
I am in version 6.9 wxdevcpp

I haven't used wxdevcpp myself, so unfortunately I can't answer questions regarding this development environment.

Regards,

Ulrich