org.apache.commons.math.dfp
Class Dfp

java.lang.Object
  extended by org.apache.commons.math.dfp.Dfp
All Implemented Interfaces:
FieldElement<Dfp>
Direct Known Subclasses:
DfpDec

public class Dfp
extends java.lang.Object
implements FieldElement<Dfp>

Decimal floating point library for Java

Another floating point class. This one is built using radix 10000 which is 104, so its almost decimal.

The design goals here are:

  1. Decimal math, or close to it
  2. Settable precision (but no mix between numbers using different settings)
  3. Portability. Code should be keep as portable as possible.
  4. Performance
  5. Accuracy - Results should always be +/- 1 ULP for basic algebraic operation
  6. Comply with IEEE 854-1987 as much as possible. (See IEEE 854-1987 notes below)

Trade offs:

  1. Memory foot print. I'm using more memory than necessary to represent numbers to get better performance.
  2. Digits are bigger, so rounding is a greater loss. So, if you really need 12 decimal digits, better use 4 base 10000 digits there can be one partially filled.

Numbers are represented in the following form:

  n  =  sign × mant × (radix)exp;

where sign is ±1, mantissa represents a fractional number between zero and one. mant[0] is the least significant digit. exp is in the range of -32767 to 32768

IEEE 854-1987 Notes and differences

IEEE 854 requires the radix to be either 2 or 10. The radix here is 10000, so that requirement is not met, but it is possible that a subclassed can be made to make it behave as a radix 10 number. It is my opinion that if it looks and behaves as a radix 10 number then it is one and that requirement would be met.

The radix of 10000 was chosen because it should be faster to operate on 4 decimal digits at once instead of one at a time. Radix 10 behavior can be realized by add an additional rounding step to ensure that the number of decimal digits represented is constant.

The IEEE standard specifically leaves out internal data encoding, so it is reasonable to conclude that such a subclass of this radix 10000 system is merely an encoding of a radix 10 system.

IEEE 854 also specifies the existence of "sub-normal" numbers. This class does not contain any such entities. The most significant radix 10000 digit is always non-zero. Instead, we support "gradual underflow" by raising the underflow flag for numbers less with exponent less than expMin, but don't flush to zero until the exponent reaches MIN_EXP-digits. Thus the smallest number we can represent would be: 1E(-(MIN_EXP-digits-1)*4), eg, for digits=5, MIN_EXP=-32767, that would be 1e-131092.

IEEE 854 defines that the implied radix point lies just to the right of the most significant digit and to the left of the remaining digits. This implementation puts the implied radix point to the left of all digits including the most significant one. The most significant digit here is the one just to the right of the radix point. This is a fine detail and is really only a matter of definition. Any side effects of this can be rendered invisible by a subclass.

Since:
2.2
Version:
$Revision: 1003889 $ $Date: 2010-10-02 23:11:55 +0200 (sam. 02 oct. 2010) $
See Also:
DfpField

Field Summary
private static java.lang.String ADD_TRAP
          Name for traps triggered by addition.
private static java.lang.String ALIGN_TRAP
          Name for traps triggered by alignment.
private static java.lang.String DIVIDE_TRAP
          Name for traps triggered by division.
static int ERR_SCALE
          The amount under/overflows are scaled by before going to trap handler
protected  int exp
          Exponent.
private  DfpField field
          Factory building similar Dfp's.
static byte FINITE
          Indicator value for normal finite numbers.
private static java.lang.String GREATER_THAN_TRAP
          Name for traps triggered by greaterThan.
static byte INFINITE
          Indicator value for Infinity.
private static java.lang.String LESS_THAN_TRAP
          Name for traps triggered by lessThan.
protected  int[] mant
          Mantissa.
static int MAX_EXP
          The maximum exponent before overflow is signaled and results flushed to infinity
static int MIN_EXP
          The minimum exponent before underflow is signaled.
private static java.lang.String MULTIPLY_TRAP
          Name for traps triggered by multiplication.
private static java.lang.String NAN_STRING
          String for NaN representation.
protected  byte nans
          Indicator for non-finite / non-number values.
private static java.lang.String NEG_INFINITY_STRING
          String for negative infinity representation.
private static java.lang.String NEW_INSTANCE_TRAP
          Name for traps triggered by newInstance.
private static java.lang.String NEXT_AFTER_TRAP
          Name for traps triggered by nextAfter.
private static java.lang.String POS_INFINITY_STRING
          String for positive infinity representation.
static byte QNAN
          Indicator value for quiet NaN.
static int RADIX
          The radix, or base of this system.
protected  byte sign
          Sign bit: & for positive, -1 for negative.
static byte SNAN
          Indicator value for signaling NaN.
private static java.lang.String SQRT_TRAP
          Name for traps triggered by square root.
private static java.lang.String TRUNC_TRAP
          Name for traps triggered by truncation.
 
Constructor Summary
  Dfp(Dfp d)
          Copy constructor.
protected Dfp(DfpField field)
          Makes an instance with a value of zero.
protected Dfp(DfpField field, byte x)
          Create an instance from a byte value.
protected Dfp(DfpField field, byte sign, byte nans)
          Creates an instance with a non-finite value.
protected Dfp(DfpField field, double x)
          Create an instance from a double value.
protected Dfp(DfpField field, int x)
          Create an instance from an int value.
protected Dfp(DfpField field, long x)
          Create an instance from a long value.
protected Dfp(DfpField field, java.lang.String s)
          Create an instance from a String representation.
 
Method Summary
 Dfp add(Dfp x)
          Add x to this.
protected  int align(int e)
          Make our exp equal to the supplied one, this may cause rounding.
 Dfp ceil()
          Round to an integer using the round ceil mode.
 int classify()
          Returns the type - one of FINITE, INFINITE, SNAN, QNAN.
private static int compare(Dfp a, Dfp b)
          Compare two instances.
protected  int complement(int extra)
          Negate the mantissa of this by computing the complement.
static Dfp copysign(Dfp x, Dfp y)
          Creates an instance that is the same as x except that it has the sign of y.
protected  java.lang.String dfp2sci()
          Convert an instance to a string using scientific notation.
protected  java.lang.String dfp2string()
          Convert an instance to a string using normal notation.
 Dfp divide(Dfp divisor)
          Divide this by divisor.
 Dfp divide(int divisor)
          Divide by a single digit less than radix.
 Dfp dotrap(int type, java.lang.String what, Dfp oper, Dfp result)
          Raises a trap.
 boolean equals(java.lang.Object other)
          Check if instance is equal to x.
 Dfp floor()
          Round to an integer using the round floor mode.
 DfpField getField()
          Get the Field (really a DfpField) to which the instance belongs.
 Dfp getOne()
          Get the constant 1.
 int getRadixDigits()
          Get the number of radix digits of the instance.
 Dfp getTwo()
          Get the constant 2.
 Dfp getZero()
          Get the constant 0.
 boolean greaterThan(Dfp x)
          Check if instance is greater than x.
 int hashCode()
          Gets a hashCode for the instance.
 int intValue()
          Convert this to an integer.
 boolean isInfinite()
          Check if instance is infinite.
 boolean isNaN()
          Check if instance is not a number.
 boolean lessThan(Dfp x)
          Check if instance is less than x.
 int log10()
          Get the exponent of the greatest power of 10 that is less than or equal to abs(this).
 int log10K()
          Get the exponent of the greatest power of 10000 that is less than or equal to the absolute value of this.
 Dfp multiply(Dfp x)
          Multiply this by x.
 Dfp multiply(int x)
          Multiply this by a single digit 0<=x<radix.
 Dfp negate()
          Returns a number that is this number with the sign bit reversed.
 Dfp newInstance()
          Create an instance with a value of 0.
 Dfp newInstance(byte x)
          Create an instance from a byte value.
 Dfp newInstance(byte sig, byte code)
          Creates an instance with a non-finite value.
 Dfp newInstance(Dfp d)
          Create an instance by copying an existing one.
 Dfp newInstance(double x)
          Create an instance from a double value.
 Dfp newInstance(int x)
          Create an instance from an int value.
 Dfp newInstance(long x)
          Create an instance from a long value.
 Dfp newInstance(java.lang.String s)
          Create an instance from a String representation.
 Dfp nextAfter(Dfp x)
          Returns the next number greater than this one in the direction of x.
 Dfp power10(int e)
          Return the specified power of 10.
 Dfp power10K(int e)
          Get the specified power of 10000.
 Dfp remainder(Dfp d)
          Returns the IEEE remainder.
 Dfp rint()
          Round to nearest integer using the round-half-even method.
protected  int round(int n)
          Round this given the next digit n using the current rounding mode.
protected  void shiftLeft()
          Shift the mantissa left, and adjust the exponent to compensate.
protected  void shiftRight()
          Shift the mantissa right, and adjust the exponent to compensate.
 Dfp sqrt()
          Compute the square root.
 Dfp subtract(Dfp x)
          Subtract x from this.
 double toDouble()
          Convert the instance into a double.
 double[] toSplitDouble()
          Convert the instance into a split double.
 java.lang.String toString()
          Get a string representation of the instance.
protected  Dfp trap(int type, java.lang.String what, Dfp oper, Dfp def, Dfp result)
          Trap handler.
protected  Dfp trunc(DfpField.RoundingMode rmode)
          Does the integer conversions with the specified rounding.
 boolean unequal(Dfp x)
          Check if instance is not equal to x.
 
Methods inherited from class java.lang.Object
clone, finalize, getClass, notify, notifyAll, wait, wait, wait
 

Field Detail

RADIX

public static final int RADIX
The radix, or base of this system. Set to 10000

See Also:
Constant Field Values

MIN_EXP

public static final int MIN_EXP
The minimum exponent before underflow is signaled. Flush to zero occurs at minExp-DIGITS

See Also:
Constant Field Values

MAX_EXP

public static final int MAX_EXP
The maximum exponent before overflow is signaled and results flushed to infinity

See Also:
Constant Field Values

ERR_SCALE

public static final int ERR_SCALE
The amount under/overflows are scaled by before going to trap handler

See Also:
Constant Field Values

FINITE

public static final byte FINITE
Indicator value for normal finite numbers.

See Also:
Constant Field Values

INFINITE

public static final byte INFINITE
Indicator value for Infinity.

See Also:
Constant Field Values

SNAN

public static final byte SNAN
Indicator value for signaling NaN.

See Also:
Constant Field Values

QNAN

public static final byte QNAN
Indicator value for quiet NaN.

See Also:
Constant Field Values

NAN_STRING

private static final java.lang.String NAN_STRING
String for NaN representation.

See Also:
Constant Field Values

POS_INFINITY_STRING

private static final java.lang.String POS_INFINITY_STRING
String for positive infinity representation.

See Also:
Constant Field Values

NEG_INFINITY_STRING

private static final java.lang.String NEG_INFINITY_STRING
String for negative infinity representation.

See Also:
Constant Field Values

ADD_TRAP

private static final java.lang.String ADD_TRAP
Name for traps triggered by addition.

See Also:
Constant Field Values

MULTIPLY_TRAP

private static final java.lang.String MULTIPLY_TRAP
Name for traps triggered by multiplication.

See Also:
Constant Field Values

DIVIDE_TRAP

private static final java.lang.String DIVIDE_TRAP
Name for traps triggered by division.

See Also:
Constant Field Values

SQRT_TRAP

private static final java.lang.String SQRT_TRAP
Name for traps triggered by square root.

See Also:
Constant Field Values

ALIGN_TRAP

private static final java.lang.String ALIGN_TRAP
Name for traps triggered by alignment.

See Also:
Constant Field Values

TRUNC_TRAP

private static final java.lang.String TRUNC_TRAP
Name for traps triggered by truncation.

See Also:
Constant Field Values

NEXT_AFTER_TRAP

private static final java.lang.String NEXT_AFTER_TRAP
Name for traps triggered by nextAfter.

See Also:
Constant Field Values

LESS_THAN_TRAP

private static final java.lang.String LESS_THAN_TRAP
Name for traps triggered by lessThan.

See Also:
Constant Field Values

GREATER_THAN_TRAP

private static final java.lang.String GREATER_THAN_TRAP
Name for traps triggered by greaterThan.

See Also:
Constant Field Values

NEW_INSTANCE_TRAP

private static final java.lang.String NEW_INSTANCE_TRAP
Name for traps triggered by newInstance.

See Also:
Constant Field Values

mant

protected int[] mant
Mantissa.


sign

protected byte sign
Sign bit: & for positive, -1 for negative.


exp

protected int exp
Exponent.


nans

protected byte nans
Indicator for non-finite / non-number values.


field

private final DfpField field
Factory building similar Dfp's.

Constructor Detail

Dfp

protected Dfp(DfpField field)
Makes an instance with a value of zero.

Parameters:
field - field to which this instance belongs

Dfp

protected Dfp(DfpField field,
              byte x)
Create an instance from a byte value.

Parameters:
field - field to which this instance belongs
x - value to convert to an instance

Dfp

protected Dfp(DfpField field,
              int x)
Create an instance from an int value.

Parameters:
field - field to which this instance belongs
x - value to convert to an instance

Dfp

protected Dfp(DfpField field,
              long x)
Create an instance from a long value.

Parameters:
field - field to which this instance belongs
x - value to convert to an instance

Dfp

protected Dfp(DfpField field,
              double x)
Create an instance from a double value.

Parameters:
field - field to which this instance belongs
x - value to convert to an instance

Dfp

public Dfp(Dfp d)
Copy constructor.

Parameters:
d - instance to copy

Dfp

protected Dfp(DfpField field,
              java.lang.String s)
Create an instance from a String representation.

Parameters:
field - field to which this instance belongs
s - string representation of the instance

Dfp

protected Dfp(DfpField field,
              byte sign,
              byte nans)
Creates an instance with a non-finite value.

Parameters:
field - field to which this instance belongs
sign - sign of the Dfp to create
nans - code of the value, must be one of INFINITE, SNAN, QNAN
Method Detail

newInstance

public Dfp newInstance()
Create an instance with a value of 0. Use this internally in preference to constructors to facilitate subclasses

Returns:
a new instance with a value of 0

newInstance

public Dfp newInstance(byte x)
Create an instance from a byte value.

Parameters:
x - value to convert to an instance
Returns:
a new instance with value x

newInstance

public Dfp newInstance(int x)
Create an instance from an int value.

Parameters:
x - value to convert to an instance
Returns:
a new instance with value x

newInstance

public Dfp newInstance(long x)
Create an instance from a long value.

Parameters:
x - value to convert to an instance
Returns:
a new instance with value x

newInstance

public Dfp newInstance(double x)
Create an instance from a double value.

Parameters:
x - value to convert to an instance
Returns:
a new instance with value x

newInstance

public Dfp newInstance(Dfp d)
Create an instance by copying an existing one. Use this internally in preference to constructors to facilitate subclasses.

Parameters:
d - instance to copy
Returns:
a new instance with the same value as d

newInstance

public Dfp newInstance(java.lang.String s)
Create an instance from a String representation. Use this internally in preference to constructors to facilitate subclasses.

Parameters:
s - string representation of the instance
Returns:
a new instance parsed from specified string

newInstance

public Dfp newInstance(byte sig,
                       byte code)
Creates an instance with a non-finite value.

Parameters:
sig - sign of the Dfp to create
code - code of the value, must be one of INFINITE, SNAN, QNAN
Returns:
a new instance with a non-finite value

getField

public DfpField getField()
Get the Field (really a DfpField) to which the instance belongs.

The field is linked to the number of digits and acts as a factory for Dfp instances.

Specified by:
getField in interface FieldElement<Dfp>
Returns:
Field (really a DfpField) to which the instance belongs

getRadixDigits

public int getRadixDigits()
Get the number of radix digits of the instance.

Returns:
number of radix digits

getZero

public Dfp getZero()
Get the constant 0.

Returns:
a Dfp with value zero

getOne

public Dfp getOne()
Get the constant 1.

Returns:
a Dfp with value one

getTwo

public Dfp getTwo()
Get the constant 2.

Returns:
a Dfp with value two

shiftLeft

protected void shiftLeft()
Shift the mantissa left, and adjust the exponent to compensate.


shiftRight

protected void shiftRight()
Shift the mantissa right, and adjust the exponent to compensate.


align

protected int align(int e)
Make our exp equal to the supplied one, this may cause rounding. Also causes de-normalized numbers. These numbers are generally dangerous because most routines assume normalized numbers. Align doesn't round, so it will return the last digit destroyed by shifting right.

Parameters:
e - desired exponent
Returns:
last digit destroyed by shifting right

lessThan

public boolean lessThan(Dfp x)
Check if instance is less than x.

Parameters:
x - number to check instance against
Returns:
true if instance is less than x and neither are NaN, false otherwise

greaterThan

public boolean greaterThan(Dfp x)
Check if instance is greater than x.

Parameters:
x - number to check instance against
Returns:
true if instance is greater than x and neither are NaN, false otherwise

isInfinite

public boolean isInfinite()
Check if instance is infinite.

Returns:
true if instance is infinite

isNaN

public boolean isNaN()
Check if instance is not a number.

Returns:
true if instance is not a number

equals

public boolean equals(java.lang.Object other)
Check if instance is equal to x.

Overrides:
equals in class java.lang.Object
Parameters:
other - object to check instance against
Returns:
true if instance is equal to x and neither are NaN, false otherwise

hashCode

public int hashCode()
Gets a hashCode for the instance.

Overrides:
hashCode in class java.lang.Object
Returns:
a hash code value for this object

unequal

public boolean unequal(Dfp x)
Check if instance is not equal to x.

Parameters:
x - number to check instance against
Returns:
true if instance is not equal to x and neither are NaN, false otherwise

compare

private static int compare(Dfp a,
                           Dfp b)
Compare two instances.

Parameters:
a - first instance in comparison
b - second instance in comparison
Returns:
-1 if ab and 0 if a==b Note this method does not properly handle NaNs or numbers with different precision.

rint

public Dfp rint()
Round to nearest integer using the round-half-even method. That is round to nearest integer unless both are equidistant. In which case round to the even one.

Returns:
rounded value

floor

public Dfp floor()
Round to an integer using the round floor mode. That is, round toward -Infinity

Returns:
rounded value

ceil

public Dfp ceil()
Round to an integer using the round ceil mode. That is, round toward +Infinity

Returns:
rounded value

remainder

public Dfp remainder(Dfp d)
Returns the IEEE remainder.

Parameters:
d - divisor
Returns:
this less n × d, where n is the integer closest to this/d

trunc

protected Dfp trunc(DfpField.RoundingMode rmode)
Does the integer conversions with the specified rounding.

Parameters:
rmode - rounding mode to use
Returns:
truncated value

intValue

public int intValue()
Convert this to an integer. If greater than 2147483647, it returns 2147483647. If less than -2147483648 it returns -2147483648.

Returns:
converted number

log10K

public int log10K()
Get the exponent of the greatest power of 10000 that is less than or equal to the absolute value of this. I.E. if this is 106 then log10K would return 1.

Returns:
integer base 10000 logarithm

power10K

public Dfp power10K(int e)
Get the specified power of 10000.

Parameters:
e - desired power
Returns:
10000e

log10

public int log10()
Get the exponent of the greatest power of 10 that is less than or equal to abs(this).

Returns:
integer base 10 logarithm

power10

public Dfp power10(int e)
Return the specified power of 10.

Parameters:
e - desired power
Returns:
10e

complement

protected int complement(int extra)
Negate the mantissa of this by computing the complement. Leaves the sign bit unchanged, used internally by add. Denormalized numbers are handled properly here.

Parameters:
extra - ???
Returns:
???

add

public Dfp add(Dfp x)
Add x to this.

Specified by:
add in interface FieldElement<Dfp>
Parameters:
x - number to add
Returns:
sum of this and x

negate

public Dfp negate()
Returns a number that is this number with the sign bit reversed.

Returns:
the opposite of this

subtract

public Dfp subtract(Dfp x)
Subtract x from this.

Specified by:
subtract in interface FieldElement<Dfp>
Parameters:
x - number to subtract
Returns:
difference of this and a

round

protected int round(int n)
Round this given the next digit n using the current rounding mode.

Parameters:
n - ???
Returns:
the IEEE flag if an exception occurred

multiply

public Dfp multiply(Dfp x)
Multiply this by x.

Specified by:
multiply in interface FieldElement<Dfp>
Parameters:
x - multiplicand
Returns:
product of this and x

multiply

public Dfp multiply(int x)
Multiply this by a single digit 0<=x<radix. There are speed advantages in this special case

Parameters:
x - multiplicand
Returns:
product of this and x

divide

public Dfp divide(Dfp divisor)
Divide this by divisor.

Specified by:
divide in interface FieldElement<Dfp>
Parameters:
divisor - divisor
Returns:
quotient of this by divisor

divide

public Dfp divide(int divisor)
Divide by a single digit less than radix. Special case, so there are speed advantages. 0 <= divisor < radix

Parameters:
divisor - divisor
Returns:
quotient of this by divisor

sqrt

public Dfp sqrt()
Compute the square root.

Returns:
square root of the instance

toString

public java.lang.String toString()
Get a string representation of the instance.

Overrides:
toString in class java.lang.Object
Returns:
string representation of the instance

dfp2sci

protected java.lang.String dfp2sci()
Convert an instance to a string using scientific notation.

Returns:
string representation of the instance in scientific notation

dfp2string

protected java.lang.String dfp2string()
Convert an instance to a string using normal notation.

Returns:
string representation of the instance in normal notation

dotrap

public Dfp dotrap(int type,
                  java.lang.String what,
                  Dfp oper,
                  Dfp result)
Raises a trap. This does not set the corresponding flag however.

Parameters:
type - the trap type
what - - name of routine trap occurred in
oper - - input operator to function
result - - the result computed prior to the trap
Returns:
The suggested return value from the trap handler

trap

protected Dfp trap(int type,
                   java.lang.String what,
                   Dfp oper,
                   Dfp def,
                   Dfp result)
Trap handler. Subclasses may override this to provide trap functionality per IEEE 854-1987.

Parameters:
type - The exception type - e.g. FLAG_OVERFLOW
what - The name of the routine we were in e.g. divide()
oper - An operand to this function if any
def - The default return value if trap not enabled
result - The result that is specified to be delivered per IEEE 854, if any
Returns:
the value that should be return by the operation triggering the trap

classify

public int classify()
Returns the type - one of FINITE, INFINITE, SNAN, QNAN.

Returns:
type of the number

copysign

public static Dfp copysign(Dfp x,
                           Dfp y)
Creates an instance that is the same as x except that it has the sign of y. abs(x) = dfp.copysign(x, dfp.one)

Parameters:
x - number to get the value from
y - number to get the sign from
Returns:
a number with the value of x and the sign of y

nextAfter

public Dfp nextAfter(Dfp x)
Returns the next number greater than this one in the direction of x. If this==x then simply returns this.

Parameters:
x - direction where to look at
Returns:
closest number next to instance in the direction of x

toDouble

public double toDouble()
Convert the instance into a double.

Returns:
a double approximating the instance
See Also:
toSplitDouble()

toSplitDouble

public double[] toSplitDouble()
Convert the instance into a split double.

Returns:
an array of two doubles which sum represent the instance
See Also:
toDouble()


Copyright (c) 2003-2013 Apache Software Foundation