FileDocCategorySizeDatePackage
SVFractionalFormat.javaAPI DocApache Poi 3.0.16548Mon Jan 01 12:39:34 GMT 2007org.apache.poi.hssf.contrib.view

SVFractionalFormat

public class SVFractionalFormat extends Format
This class is used to format cells into their fractional format. I cant be 100% sure that the same fractional value will be displayed as in excel but then again it is a lossy formating mode anyway
author
Jason Height
since
15 July 2002

Fields Summary
private short
ONE_DIGIT
private short
TWO_DIGIT
private short
THREE_DIGIT
private short
UNITS
private int
units
private short
mode
Constructors Summary
public SVFractionalFormat(String formatStr)
Constructs a new FractionalFormatter The formatStr defines how the number will be formatted # ?/? Up to one digit # ??/?? Up to two digits # ???/??? Up to three digits # ?/2 In halves # ?/4 In quarters # ?/8 In eighths # ?/16 In sixteenths # ?/10 In tenths # ?/100 In hundredths


                                                                      
     
    if ("# ?/?".equals(formatStr))
      mode = ONE_DIGIT;
    else if ("# ??/??".equals(formatStr))
      mode = TWO_DIGIT;
    else if ("# ???/???".equals(formatStr))
      mode = THREE_DIGIT;
    else if ("# ?/2".equals(formatStr)) {
      mode = UNITS;
      units = 2;
    } else if ("# ?/4".equals(formatStr)) {
      mode = UNITS;
      units = 4;
    } else if ("# ?/8".equals(formatStr)) {
      mode = UNITS;
      units = 8;
    } else if ("# ?/16".equals(formatStr)) {
      mode = UNITS;
      units = 16;
    } else if ("# ?/10".equals(formatStr)) {
      mode = UNITS;
      units = 10;
    } else if ("# ?/100".equals(formatStr)) {
      mode = UNITS;
      units = 100;
    }
Methods Summary
public java.lang.Objectclone()

    //JMH TBD
    return null;
private java.lang.Stringformat(double f, int MaxDen)
Returns a fractional string representation of a double to a maximum denominator size This code has been translated to java from the following web page. http://www.codeproject.com/cpp/fraction.asp Originally coded in c++ By Dean Wyant dwyant@mindspring.com The code on the web page is freely available.

param
f Description of the Parameter
param
MaxDen Description of the Parameter
return
Description of the Return Value

    long Whole = (long)f;
    int sign = 1;
    if (f < 0) {
      sign = -1;
    }
    double Precision = 0.00001;
    double AllowedError = Precision;
    double d = Math.abs(f);
    d -= Whole;
    double Frac = d;
    double Diff = Frac;
    long Num = 1;
    long Den = 0;
    long A = 0;
    long B = 0;
    long i = 0;
    if (Frac > Precision) {
      while (true) {
        d = 1.0 / d;
        i = (long) (d + Precision);
        d -= i;
        if (A > 0) {
          Num = i * Num + B;
        }
        Den = (long) (Num / Frac + 0.5);
        Diff = Math.abs((double) Num / Den - Frac);
        if (Den > MaxDen) {
          if (A > 0) {
            Num = A;
            Den = (long) (Num / Frac + 0.5);
            Diff = Math.abs((double) Num / Den - Frac);
          } else {
            Den = MaxDen;
            Num = 1;
            Diff = Math.abs((double) Num / Den - Frac);
            if (Diff > Frac) {
              Num = 0;
              Den = 1;
              // Keeps final check below from adding 1 and keeps Den from being 0
              Diff = Frac;
            }
          }
          break;
        }
        if ((Diff <= AllowedError) || (d < Precision)) {
          break;
        }
        Precision = AllowedError / Diff;
        // This calcualtion of Precision does not always provide results within
        // Allowed Error. It compensates for loss of significant digits that occurs.
        // It helps to round the inprecise reciprocal values to i.
        B = A;
        A = Num;
      }
    }
    if (Num == Den) {
      Whole++;
      Num = 0;
      Den = 0;
    } else if (Den == 0) {
      Num = 0;
    }
    if (sign < 0) {
      if (Whole == 0) {
        Num = -Num;
      } else {
        Whole = -Whole;
      }
    }
    return new StringBuffer().append(Whole).append(" ").append(Num).append("/").append(Den).toString();
public final java.lang.Stringformat(double val)

    if (mode == ONE_DIGIT) {
      return format(val, 9);
    } else if (mode == TWO_DIGIT) {
      return format(val, 99);
    } else if (mode == THREE_DIGIT) {
      return format(val, 999);
    } else if (mode == UNITS) {
      return formatUnit(val , units);
    }
    throw new RuntimeException("Unexpected Case");
public java.lang.StringBufferformat(java.lang.Object obj, java.lang.StringBuffer toAppendTo, java.text.FieldPosition pos)

    if (obj instanceof Number) {
      toAppendTo.append(format(((Number)obj).doubleValue()));
      return toAppendTo;
    }
    else throw new IllegalArgumentException("Can only handle Numbers");
private java.lang.StringformatUnit(double f, int units)
This method formats the double in the units specified. The usints could be any number but in this current implementation it is halves (2), quaters (4), eigths (8) etc

    long Whole = (long)f;
    f -= Whole;
    long Num = Math.round(f * units);

    return new StringBuffer().append(Whole).append(" ").append(Num).append("/").append(units).toString();
public java.lang.ObjectparseObject(java.lang.String source, java.text.ParsePosition status)

    //JMH TBD
    return null;
public java.lang.ObjectparseObject(java.lang.String source)

    //JMH TBD
    return null;