public final class UncheckedScale0fTruncatingArithmetic extends AbstractUncheckedScale0fArithmetic
Scale0f, that is, for longs. If an operation leads to an overflow the
result is silently truncated.| Modifier and Type | Field and Description |
|---|---|
static UncheckedScale0fTruncatingArithmetic |
INSTANCE
The singleton instance.
|
| Constructor and Description |
|---|
UncheckedScale0fTruncatingArithmetic() |
| Modifier and Type | Method and Description |
|---|---|
long |
addUnscaled(long uDecimal,
long unscaled,
int scale)
Returns an unscaled decimal whose value is
(uDecimal + unscaled * 10-scale). |
long |
avg(long a,
long b)
Returns an unscaled decimal whose value is the average of
uDecimal1 and uDecimal2. |
long |
divide(long uDecimalDividend,
long uDecimalDivisor)
Returns an unscaled decimal whose value is the quotient of the specified arguments:
(uDecimalDividend / uDecimalDivisor). |
long |
divideByLong(long uDecimalDividend,
long lDivisor)
Returns an unscaled decimal whose value is
(uDecimalDividend / lDivisor) where the second argument is a
true long value instead of an unscaled decimal. |
long |
divideByPowerOf10(long uDecimal,
int positions)
Returns an unscaled decimal whose value is
(uDecimal / 10n). |
long |
divideByUnscaled(long uDecimal,
long unscaled,
int scale)
Returns an unscaled decimal whose value is
(uDecimal / (unscaled * 10-scale)). |
long |
fromBigDecimal(BigDecimal value)
Converts the specified
BigDecimal value to an unscaled decimal. |
long |
fromDouble(double value)
Converts the specified double value to an unscaled decimal.
|
long |
fromFloat(float value)
Converts the specified float value to an unscaled decimal.
|
long |
fromUnscaled(long unscaledValue,
int scale)
Converts the specified unscaled decimal with the given scale to another unscaled decimal of the scale of this
arithmetic.
|
RoundingMode |
getRoundingMode()
Returns the rounding mode applied to operations of this
DecimalArithmetic object if rounding is
necessary. |
UncheckedRounding |
getTruncationPolicy()
Returns the truncation policy defining how to handle truncation due to overflow or rounding.
|
long |
invert(long uDecimal)
Returns an unscaled decimal whose value is the inverse of the argument:
1/uDecimal. |
long |
multiplyByPowerOf10(long uDecimal,
int positions)
Returns an unscaled decimal whose value is
(uDecimal * 10n). |
long |
multiplyByUnscaled(long uDecimal,
long unscaled,
int scale)
Returns an unscaled decimal whose value is
(uDecimal * unscaled * 10-scale). |
long |
parse(CharSequence value,
int start,
int end)
Translates the string representation of a
Decimal into an unscaled Decimal. |
long |
parse(String value)
Translates the string representation of a
Decimal into an unscaled Decimal. |
long |
pow(long uDecimal,
int exponent)
Returns an unscaled decimal whose value is
(uDecimalBaseexponent). |
long |
round(long uDecimal,
int precision)
Returns an unscaled decimal whose value is rounded to the specified
precision using the
rounding mode of this arithmetic. |
long |
shiftLeft(long uDecimal,
int positions)
Returns an unscaled decimal whose value is
(uDecimal << n). |
long |
shiftRight(long uDecimal,
int positions)
Returns an unscaled decimal whose value is
(uDecimal >> n). |
long |
sqrt(long uDecimal)
Returns an unscaled decimal whose value is the square root of the specified argument:
sqrt(uDecimal). |
long |
subtractUnscaled(long uDecimal,
long unscaled,
int scale)
Returns an unscaled decimal whose value is
(uDecimal - unscaled * 10-scale). |
double |
toDouble(long uDecimal)
Converts the specified unscaled decimal value into a double value and returns it.
|
float |
toFloat(long uDecimal)
Converts the specified unscaled decimal value into a float value and returns it.
|
long |
toUnscaled(long uDecimal,
int scale)
Converts the specified unscaled decimal value into an unscaled value of the given scale.
|
addLong, fromLong, getScale, getScaleMetrics, multiply, one, square, subtractLong, toLong, toString, toStringabs, add, getOverflowMode, multiplyByLong, negate, subtractcompare, compareToUnscaled, deriveArithmetic, deriveArithmetic, deriveArithmetic, deriveArithmetic, deriveArithmetic, fromBigInteger, signum, toBigDecimal, toBigDecimalpublic static final UncheckedScale0fTruncatingArithmetic INSTANCE
public UncheckedScale0fTruncatingArithmetic()
public final RoundingMode getRoundingMode()
DecimalArithmeticDecimalArithmetic object if rounding is
necessary.DecimalArithmetic object if rounding is necessarypublic final UncheckedRounding getTruncationPolicy()
DecimalArithmeticTruncationPolicy is defined by the overflow mode and the
rounding mode.getTruncationPolicy in interface DecimalArithmeticgetTruncationPolicy in class AbstractUncheckedArithmeticDecimalArithmetic handles truncationpublic final long addUnscaled(long uDecimal, long unscaled, int scale)
DecimalArithmetic(uDecimal + unscaled * 10-scale). If rounding must
be performed, this arithmetic's rounding mode is applied. Note that scale of the first
operand is determined by this arithmetic's scale whereas the scale of the second
unscaled value is explicitly specified by the scale argument.
Mathematically the method calculates round(uDecimal + lValue * 10-scale + s) where s
refers to this arithetic's scale. The method avoids information loss due to overflow of intermediary results.
uDecimal - unscaled decimal value to be added tounscaled - unscaled value to addscale - scale associated with the unscaled valueround(uDecimal + unscaled * 10-scale)public final long subtractUnscaled(long uDecimal, long unscaled, int scale)
DecimalArithmetic(uDecimal - unscaled * 10-scale). If rounding must
be performed, this arithmetic's rounding mode is applied. Note that scale of the first
operand is determined by this arithmetic's scale whereas the scale of the second
unscaled value is explicitly specified by the scale argument.
Mathematically the method calculates round(uDecimal - lValue * 10-scale + s) where s
refers to this arithetic's scale. The method avoids information loss due to overflow of intermediary results.
uDecimal - unscaled decimal value to subtract fromunscaled - unscaled value to subtractscale - scale associated with the unscaled valueround(uDecimal - unscaled * 10-scale) as unscaled decimalpublic final long multiplyByUnscaled(long uDecimal, long unscaled, int scale)
DecimalArithmetic(uDecimal * unscaled * 10-scale). If rounding must
be performed, this arithmetic's rounding mode is applied. Note that scale of the first
operand is determined by this arithmetic's scale whereas the scale of the second
unscaled value is explicitly specified by the scale argument.
Mathematically the method calculates round((uDecimal * unscaled) * 10-scale) avoiding
information loss due to overflow of intermediary results.
uDecimal - unscaled decimal value to be multipliedunscaled - unscaled value to be multipliedscale - scale associated with the unscaled valueround(uDecimal * (unscaled * 10-scale))public final long divide(long uDecimalDividend, long uDecimalDivisor)
DecimalArithmetic(uDecimalDividend / uDecimalDivisor). If rounding must be performed, this arithmetic's
rounding mode is applied.
Mathematically the method calculates round((uDecimalDividend * 10scale) / uDecimalDivisor)
avoiding information loss due to overflow of intermediary results.
uDecimalDividend - value to be divided.uDecimalDivisor - value by which the dividend is to be divided.round(uDecimalDividend / uDecimalDivisor)public final long divideByLong(long uDecimalDividend, long lDivisor)
DecimalArithmetic(uDecimalDividend / lDivisor) where the second argument is a
true long value instead of an unscaled decimal. If rounding must be performed, this arithmetic's
rounding mode is applied.uDecimalDividend - value to be divided.lDivisor - long value by which the dividend is to be divided.round(uDecimalDividend / lDivisor)public final long divideByUnscaled(long uDecimal, long unscaled, int scale)
DecimalArithmetic(uDecimal / (unscaled * 10-scale)). If rounding
must be performed, this arithmetic's rounding mode is applied. Note that scale of the
first operand is determined by this arithmetic's scale whereas the scale of the second
unscaled value is explicitly specified by the scale argument.
Mathematically the method calculates round((uDecimal * 10scale) / unscaled) avoiding
information loss due to overflow of intermediary results.
uDecimal - value to be divided.unscaled - unscaled value by which the dividend is to be divided.scale - scale associated with the unscaled valueround(uDecimal / (unscaled * 10-scale))public final long multiplyByPowerOf10(long uDecimal, int positions)
DecimalArithmetic(uDecimal * 10n).
The power, n, may be negative, in which case this method performs a division by a power of ten. If
rounding must be performed (for negative n), this arithmetic's rounding mode is
applied.
uDecimal - value to be multipliedpositions - the power of tenround(uDecimal × 10n)public final long divideByPowerOf10(long uDecimal, int positions)
DecimalArithmetic(uDecimal / 10n). If rounding must be performed,
this arithmetic's rounding mode is applied.
The power, n, may be negative, in which case this method performs a multiplication by a power of ten.
uDecimal - value to be divided.positions - the power of tenround(uDecimal / 10n)public final long invert(long uDecimal)
DecimalArithmetic1/uDecimal. If rounding must be
performed, this arithmetic's rounding mode is applied.
Mathematically the method calculates round((10scale * 10scale) / uDecimalDivisor)
avoiding information loss due to overflow of intermediary results.
uDecimal - the unscaled decimal value to invertround(1/uDecimal)public final long sqrt(long uDecimal)
DecimalArithmeticsqrt(uDecimal). If
rounding must be performed, this arithmetic's rounding mode is applied.
Mathematically the method calculates round(sqrt(uDecimal * 10scale)) avoiding information
loss due to overflow of intermediary results.
uDecimal - the unscaled decimal valueround(sqrt(uDecimal))public final long pow(long uDecimal, int exponent)
DecimalArithmetic(uDecimalBaseexponent). Note that exponent
is an integer rather than a decimal. If rounding must be performed, this arithmetic's rounding mode is applied.
The current implementation uses the core algorithm defined in ANSI standard X3.274-1996. For exponent >= 0, the
returned numerical value is within 1 ULP of the exact numerical value; the result is actually exact for all
rounding modes other than HALF_UP, HALF_EVEN and HALF_DOWN. No precision is guaranteed for exponent < 0 but the
result is usually exact up to 10-20 ULP.
Properties of the X3.274-1996 algorithm are:
IllegalArgumentException is thrown if abs(n) > 999999999exponent is zero, one is returned even if uDecimalBase is zero, otherwise
exponent is positive, the result is calculated via the repeated squaring technique into a single
accumulatorexponent is negative, the result is calculated as if exponent were positive; this value is then divided
into onerounding mode
Note: this operation is not strictly garbage free since internally, two ThreadLocal objects are
used to calculate the result. The ThreadLocal values may become garbage if the thread becomes garbage.
uDecimal - the unscaled decimal base valueexponent - exponent to which uDecimalBase is to be raised.uDecimalBaseexponentpublic final long shiftLeft(long uDecimal, int positions)
DecimalArithmetic(uDecimal << n). The shift distance, n, may be
negative, in which case this method performs a right shift. The result is equal to
round(uDecimal * 2n) using this arithmetic's rounding mode if
rounding is necessary.uDecimal - the unscaled decimal value to shiftpositions - shift distance, in bits.round(uDecimal << n)DecimalArithmetic.shiftRight(long, int)public final long shiftRight(long uDecimal, int positions)
DecimalArithmetic(uDecimal >> n). The shift distance, n, may be
negative, in which case this method performs a left shift. The result is equal to
round(uDecimal / 2n) using this arithmetic's rounding mode if
rounding is necessary.uDecimal - the unscaled decimal value to shiftpositions - shift distance, in bits.round(uDecimal >> n)DecimalArithmetic.shiftLeft(long, int)public final long avg(long a, long b)
DecimalArithmeticuDecimal1 and uDecimal2. The method is
much more efficient than an addition and subsequent long division and is guaranteed not to overflow. If rounding
must be performed, this arithmetic's rounding mode is applied.a - the first unscaled decimal value to averageb - the second unscaled decimal value to averageround((uDecimal1 + uDecimal2) / 2)public final long round(long uDecimal, int precision)
DecimalArithmeticprecision using the
rounding mode of this arithmetic.
Note that this method does not change the scale of the value --- extra digits are simply zeroised.
Examples and special cases:
IllegalArgumentException is thrownuDecimal - the unscaled decimal value to roundprecision - the precision to use for the rounding, for instance 2 to round to the second digit after the decimal
point; must be at least (scale - 18)public final long toUnscaled(long uDecimal, int scale)
DecimalArithmeticrounding mode is applied if rounding is necessary.uDecimal - the unscaled decimal value to convert into an unscaled value of the specified scalescale - the target scale for the result valueuDecimal value converted into an unscaled value of the specified scale, possibly rounded or
truncatedpublic final double toDouble(long uDecimal)
DecimalArithmeticrounding mode is applied if rounding is necessary.uDecimal - the unscaled decimal value to convert into a double valueuDecimal value converted into a double value, possibly rounded or truncatedpublic final float toFloat(long uDecimal)
DecimalArithmeticrounding mode is applied if rounding is necessary.uDecimal - the unscaled decimal value to convert into a float valueuDecimal value converted into a float value, possibly rounded or truncatedpublic final long fromUnscaled(long unscaledValue, int scale)
DecimalArithmeticunscaledValue - the unscaled decimal value to convertscale - the scale associated with unscaledValuepublic final long fromFloat(float value)
DecimalArithmeticscale.value - the value to convertpublic final long fromDouble(double value)
DecimalArithmeticscale.value - the value to convertpublic final long fromBigDecimal(BigDecimal value)
DecimalArithmeticBigDecimal value to an unscaled decimal. An exception is thrown if the specified
value is too large to be represented as a Decimal of this arithmetic's scale.
Note: this operation is not garbage free, meaning that new temporary objects may be allocated during the conversion.
value - the value to convertpublic final long parse(String value)
DecimalArithmeticDecimal into an unscaled Decimal. The string representation
consists of an optional sign, '+' or '-' , followed by a sequence of zero or more decimal digits
("the integer"), optionally followed by a fraction.
The fraction consists of a decimal point followed by zero or more decimal digits. The string must contain at
least one digit in either the integer or the fraction. If the fraction contains more digits than this
arithmetic's scale, the value is rounded using the arithmetic's rounding mode. An exception is thrown if the value is too large to be represented as a Decimal of this
arithmetic's scale.
value - a String containing the decimal value representation to be parsedlong valuepublic final long parse(CharSequence value, int start, int end)
DecimalArithmeticDecimal into an unscaled Decimal. The string representation
consists of an optional sign, '+' or '-' , followed by a sequence of zero or more decimal digits
("the integer"), optionally followed by a fraction.
The fraction consists of a decimal point followed by zero or more decimal digits. The string must contain at
least one digit in either the integer or the fraction. If the fraction contains more digits than this
arithmetic's scale, the value is rounded using the arithmetic's rounding mode. An exception is thrown if the value is too large to be represented as a Decimal of this
arithmetic's scale.
value - a character sequence such as a String containing the decimal value representation to be parsedstart - the start index to read characters in value, inclusiveend - the end index where to stop reading in characters in value, exclusivelong value