net.sf.saxon.value
Class NumericValue

java.lang.Object
  net.sf.saxon.value.Value
      net.sf.saxon.value.AtomicValue
          net.sf.saxon.value.NumericValue

All Implemented Interfaces:

Serializable, Comparable, Expression, Item

Direct Known Subclasses:

DecimalValue, DoubleValue, FloatValue, IntegerValue

public abstract class NumericValue
extends AtomicValue
implements Comparable

NumericValue is an abstract superclass for IntegerValue, DecimalValue, FloatValue, and DoubleValue

See Also:

Serialized Form

Constructor Summary

NumericValue()

Method Summary

abstract NumericValue	arithmetic(int operator, NumericValue other) Perform a binary arithmetic operation
double	asDouble() Get the numeric value as a double
long	asLong() Return the numeric value as a Java long.
abstract NumericValue	ceiling() Implement the XPath ceiling() function
int	compareTo(Object other) Compare the value to another numeric value
boolean	equals(Object other) The equals() function compares numeric equality among integers, decimals, floats, doubles, and their subtypes
abstract NumericValue	floor() Implement the XPath floor() function
abstract int	hashCode() hashCode() must be the same for two values that are equal.
boolean	isNaN() Test whether the value is the double/float value NaN
abstract boolean	isWholeNumber() Determine whether the value is a whole number, that is, whether it compares equal to some integer
abstract NumericValue	negate() Change the sign of the number
static ItemType	promote(ItemType v1, ItemType v2) Identify lowest common supertype of two numeric values for promotion purposes
abstract NumericValue	round() Implement the XPath round() function
abstract NumericValue	roundToHalfEven(int scale) Implement the XPath 2.0 round-to-half-even() function
String	toString() Produce a string representation of the value

Methods inherited from class net.sf.saxon.value.AtomicValue

conversionPreference, convert, convert, convert, convertToJava, effectiveBooleanValue, evaluateAsString, evaluateItem, getCardinality, getComponent, getPrimitiveValue, getTypedValue, iterate

Methods inherited from class net.sf.saxon.value.Value

analyze, asItem, asValue, collapseWhitespace, getDependencies, getSpecialProperties, getSubExpressions, inverse, normalizeWhitespace, promote, simplify, stringToInteger, stringToNumber, trimWhitespace

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Methods inherited from interface net.sf.saxon.om.Item

getStringValue

Methods inherited from interface net.sf.saxon.expr.Expression

display, getItemType

Constructor Detail

NumericValue

public NumericValue()

Method Detail

asDouble

public double asDouble()

Get the numeric value as a double

Returns:

A double representing this numeric value; NaN if it cannot be converted

isNaN

public boolean isNaN()

Test whether the value is the double/float value NaN

asLong

public long asLong()
            throws XPathException

Return the numeric value as a Java long.

Returns:

the numeric value as a Java long. This performs truncation towards zero.

Throws:

XPathException - if the value cannot be converted

negate

public abstract NumericValue negate()
                             throws XPathException

Change the sign of the number

Returns:

a value, of the same type as the original, with its sign inverted

Throws:

XPathException - if the conversion is not possible

floor

public abstract NumericValue floor()

Implement the XPath floor() function

Returns:

a value, of the same type as that supplied, rounded towards minus infinity

ceiling

public abstract NumericValue ceiling()

Implement the XPath ceiling() function

Returns:

a value, of the same type as that supplied, rounded towards plus infinity

round

public abstract NumericValue round()

Implement the XPath round() function

Returns:

a value, of the same type as that supplied, rounded towards the nearest whole number (0.5 rounded up)

roundToHalfEven

public abstract NumericValue roundToHalfEven(int scale)

Implement the XPath 2.0 round-to-half-even() function

Parameters:

scale - the decimal position for rounding: e.g. 2 rounds to a multiple of 0.01, while -2 rounds to a multiple of 100

Returns:

a value, of the same type as the original, rounded towards the nearest multiple of 10**(-scale), with rounding towards the nearest even number if two values are equally near

arithmetic

public abstract NumericValue arithmetic(int operator,
                                        NumericValue other)
                                 throws XPathException

Perform a binary arithmetic operation

Parameters:

operator - the binary arithmetic operation to be performed. Uses the constants defined in the Tokenizer class

other - the other operand

Returns:

the result of the arithmetic operation

Throws:

XPathException - if an arithmetic error occurs

See Also:

Tokenizer

isWholeNumber

public abstract boolean isWholeNumber()

Determine whether the value is a whole number, that is, whether it compares equal to some integer

Returns:

true if the value is a whole number

compareTo

public int compareTo(Object other)

Compare the value to another numeric value

Specified by:

compareTo in interface Comparable

Parameters:

other - The other numeric value

Returns:

-1 if this one is the lower, 0 if they are numerically equal, +1 if this one is the higher.

Throws:

ClassCastException - if the other value is not a NumericValue (the parameter is declared as Object to satisfy the Comparable interface)

equals

public final boolean equals(Object other)

The equals() function compares numeric equality among integers, decimals, floats, doubles, and their subtypes

Overrides:

equals in class Object

Parameters:

other - the value to be compared with this one

Returns:

true if the two values are numerically equal

promote

public static ItemType promote(ItemType v1,
                               ItemType v2)

Identify lowest common supertype of two numeric values for promotion purposes

Parameters:

v1 - the item type of the first operand

v2 - the item type of the second operand

Returns:

the item type that should be used for arithmetic between operands of the two specified item types

hashCode

public abstract int hashCode()

hashCode() must be the same for two values that are equal. One way to ensure this is to convert the value to a double, and take the hashCode of the double. But this is expensive in the common case where we are comparing integers. So we adopt the rule: for values that are in the range of a Java Integer, we use the int value as the hashcode. For values outside that range, we convert to a double and take the hashCode of the double. This method needs to have a compatible implementation in each subclass.

Overrides:

hashCode in class Object

Returns:

the hash code of the numeric value

toString

public String toString()

Produce a string representation of the value

Overrides:

toString in class Object

Returns:

The result of casting the number to a string