com.fasterxml.jackson.core.io.schubfach

Class FloatToDecimal

java.lang.Object

com.fasterxml.jackson.core.io.schubfach.FloatToDecimal

public final class FloatToDecimal
extends Object

This class exposes a method to render a float as a string.

Author:

Raffaello Giulietti

Field Summary

Fields

Modifier and Type	Field and Description
int	MAX_CHARS

Method Summary

Modifier and Type	Method and Description
static Appendable	appendTo(float v, Appendable app) Appends the rendering of the v to app.
static String	toString(float v) Returns a string rendering of the float argument.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

MAX_CHARS

public final int MAX_CHARS

See Also:

Constant Field Values

Method Detail

toString

public static String toString(float v)

Returns a string rendering of the float argument.

The characters of the result are all drawn from the ASCII set.

• Any NaN, whether quiet or signaling, is rendered as "NaN", regardless of the sign bit.
• The infinities +∞ and -∞ are rendered as "Infinity" and "-Infinity", respectively.
• The positive and negative zeroes are rendered as "0.0" and "-0.0", respectively.
• A finite negative v is rendered as the sign '-' followed by the rendering of the magnitude -v.
• A finite positive v is rendered in two stages:
  • Selection of a decimal: A well-defined decimal d_v is selected to represent v.
  • Formatting as a string: The decimal d_v is formatted as a string, either in plain or in computerized scientific notation, depending on its value.

A decimal is a number of the form d×10ⁱ for some (unique) integers d > 0 and i such that d is not a multiple of 10. These integers are the significand and the exponent, respectively, of the decimal. The length of the decimal is the (unique) integer n meeting 10^n-1 ≤ d < 10ⁿ.

The decimal d_v for a finite positive v is defined as follows:

• Let R be the set of all decimals that round to v according to the usual round-to-closest rule of IEEE 754 floating-point arithmetic.
• Let m be the minimal length over all decimals in R.
• When m ≥ 2, let T be the set of all decimals in R with length m. Otherwise, let T be the set of all decimals in R with length 1 or 2.
• Define d_v as the decimal in T that is closest to v. Or if there are two such decimals in T, select the one with the even significand (there is exactly one).

The (uniquely) selected decimal d_v is then formatted.

Let d, i and n be the significand, exponent and length of d_v, respectively. Further, let e = n + i - 1 and let d₁…d_n be the usual decimal expansion of the significand. Note that d₁ ≠ 0 ≠ d_n.

• Case -3 ≤ e < 0: d_v is formatted as 0.0…0d₁…d_n, where there are exactly -(n + i) zeroes between the decimal point and d₁. For example, 123 × 10^-4 is formatted as 0.0123.
• Case 0 ≤ e < 7:
  • Subcase i ≥ 0: d_v is formatted as d₁…d_n0…0.0, where there are exactly i zeroes between d_n and the decimal point. For example, 123 × 10² is formatted as 12300.0.
  • Subcase i < 0: d_v is formatted as d₁…d_n+i.d_n+i+1…d_n. There are exactly -i digits to the right of the decimal point. For example, 123 × 10^-1 is formatted as 12.3.
• Case e < -3 or e ≥ 7: computerized scientific notation is used to format d_v. Here e is formatted as by Integer.toString(int).
  • Subcase n = 1: d_v is formatted as d₁.0Ee. For example, 1 × 10²³ is formatted as 1.0E23.
  • Subcase n > 1: d_v is formatted as d₁.d₂…d_nEe. For example, 123 × 10^-21 is formatted as 1.23E-19.

Parameters:

v - the float to be rendered.

Returns:

a string rendering of the argument.

appendTo

public static Appendable appendTo(float v,
                                  Appendable app)
                           throws IOException

Appends the rendering of the v to app.

The outcome is the same as if v were first rendered and the resulting string were then appended to app.

Parameters:

v - the float whose rendering is appended.

app - the Appendable to append to.

Throws:

IOException - If an I/O error occurs