org.unitConverter.math

Class DecimalComparison

java.lang.Object

org.unitConverter.math.DecimalComparison

public final class DecimalComparison
extends java.lang.Object

A class that contains methods to compare float and double values.

Since:

2019-03-18, v0.2.0

Author:

Adrien Hopkins

Field Summary

Fields

Modifier and Type	Field and Description
static double	DOUBLE_EPSILON The value used for double comparison.
static float	FLOAT_EPSILON The value used for float comparison.

Method Summary

Modifier and Type	Method and Description
static boolean	equals(double a, double b) Tests for equality of double values using DOUBLE_EPSILON.
static boolean	equals(double a, double b, double epsilon) Tests for double equality using a custom epsilon value.
static boolean	equals(float a, float b) Tests for equality of float values using FLOAT_EPSILON.
static boolean	equals(float a, float b, float epsilon) Tests for float equality using a custom epsilon value.
static int	hash(double d) Takes the hash code of doubles.

Methods inherited from class java.lang.Object

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

Field Detail

DOUBLE_EPSILON

public static final double DOUBLE_EPSILON

The value used for double comparison. If two double values are within this value multiplied by the larger value, they are considered equal.

Since:

2019-03-18, v0.2.0

See Also:

Constant Field Values

FLOAT_EPSILON

public static final float FLOAT_EPSILON

The value used for float comparison. If two float values are within this value multiplied by the larger value, they are considered equal.

Since:

2019-03-18, v0.2.0

See Also:

Constant Field Values

Method Detail

equals

public static final boolean equals(double a,
                                   double b)

Tests for equality of double values using DOUBLE_EPSILON.

WARNING: this method is not technically transitive. If a and b are off by slightly less than epsilon * max(abs(a), abs(b)), and b and c are off by slightly less than epsilon * max(abs(b), abs(c)), then equals(a, b) and equals(b, c) will both return true, but equals(a, c) will return false. However, this situation is very unlikely to ever happen in a real programming situation.

If this does become a concern, some ways to solve this problem:

1. Raise the value of epsilon using equals(double, double, double) (this does not make a violation of transitivity impossible, it just significantly reduces the chances of it happening)
2. Use BigDecimal instead of double (this will make a violation of transitivity 100% impossible)

Parameters:

a - first value to test

b - second value to test

Returns:

whether they are equal

Since:

2019-03-18, v0.2.0

See Also:

#hashCode(double)

equals

public static final boolean equals(double a,
                                   double b,
                                   double epsilon)

Tests for double equality using a custom epsilon value.

WARNING: this method is not technically transitive. If a and b are off by slightly less than epsilon * max(abs(a), abs(b)), and b and c are off by slightly less than epsilon * max(abs(b), abs(c)), then equals(a, b) and equals(b, c) will both return true, but equals(a, c) will return false. However, this situation is very unlikely to ever happen in a real programming situation.

If this does become a concern, some ways to solve this problem:

1. Raise the value of epsilon (this does not make a violation of transitivity impossible, it just significantly reduces the chances of it happening)
2. Use BigDecimal instead of double (this will make a violation of transitivity 100% impossible)

Parameters:

a - first value to test

b - second value to test

epsilon - allowed difference

Returns:

whether they are equal

Since:

2019-03-18, v0.2.0

equals

public static final boolean equals(float a,
                                   float b)

Tests for equality of float values using FLOAT_EPSILON.

WARNING: this method is not technically transitive. If a and b are off by slightly less than epsilon * max(abs(a), abs(b)), and b and c are off by slightly less than epsilon * max(abs(b), abs(c)), then equals(a, b) and equals(b, c) will both return true, but equals(a, c) will return false. However, this situation is very unlikely to ever happen in a real programming situation.

If this does become a concern, some ways to solve this problem:

1. Raise the value of epsilon using equals(float, float, float) (this does not make a violation of transitivity impossible, it just significantly reduces the chances of it happening)
2. Use BigDecimal instead of float (this will make a violation of transitivity 100% impossible)

Parameters:

a - first value to test

b - second value to test

Returns:

whether they are equal

Since:

2019-03-18, v0.2.0

equals

public static final boolean equals(float a,
                                   float b,
                                   float epsilon)

Tests for float equality using a custom epsilon value.

WARNING: this method is not technically transitive. If a and b are off by slightly less than epsilon * max(abs(a), abs(b)), and b and c are off by slightly less than epsilon * max(abs(b), abs(c)), then equals(a, b) and equals(b, c) will both return true, but equals(a, c) will return false. However, this situation is very unlikely to ever happen in a real programming situation.

If this does become a concern, some ways to solve this problem:

1. Raise the value of epsilon (this does not make a violation of transitivity impossible, it just significantly reduces the chances of it happening)
2. Use BigDecimal instead of float (this will make a violation of transitivity 100% impossible)

Parameters:

a - first value to test

b - second value to test

epsilon - allowed difference

Returns:

whether they are equal

Since:

2019-03-18, v0.2.0

hash

public static final int hash(double d)

Takes the hash code of doubles. Values that are equal according to equals(double, double) will have the same hash code.

Parameters:

d - double to hash

Returns:

hash code of double

Since:

2019-10-16