com.pervasive.datarush.util

Class HashUtil

java.lang.Object

com.pervasive.datarush.util.HashUtil

public final class HashUtil
extends Object

Utility methods for hashing different data types.

These hashes are based on C code placed in the public domain by Bob Jenkins. They are advertised as strong enough to allow the use of hash & mask instead of hash % prime for computing a hash bucket.

Field Summary

Fields

Modifier and Type	Field and Description
static int	NULL_HASH_CODE An arbitrary nonzero constant for the hash code of null valued tokens.

Constructor Summary

Constructors

Constructor and Description
HashUtil()

Method Summary

Modifier and Type	Method and Description
static int	binaryHashCode(byte[] bytes) Hash a binary value using Bob Jenkins' "One-at-a-Time Hash".
static int	binaryHashCode(byte[] bytes, int offset, int len) Hash a binary value using Bob Jenkins' "One-at-a-Time Hash".
static int	booleanHashCode(boolean b) Hashes a boolean value using the same algorithm as intHashCode(int).
static int	charHashCode(char c) Hashes a character value using the same algorithm as intHashCode(int).
static int	doubleHashCode(double d) Hashes a double value using the same algorithm as longHashCode(long).
static int	finalMix(int a, int b, int c) Replicates the computation of Bob Jenkins' "final" macro.
static int	floatHashCode(float f) Hashes a float value using the same algorithm as intHashCode(int).
static int	intHashCode(int a) Hash a binary value using Bob Jenkins' "4-byte Integer Hashing".
static int	longHashCode(long l) Hashes a long value.
static int	objectHashCode(Object o) Hashes an object using Object.hashCode().
static int	timestampHashCode(long epochSecs, int nanos) Hashes a timestamp decomposed in seconds and nanos parts.

Methods inherited from class java.lang.Object

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

Field Detail

NULL_HASH_CODE

public static final int NULL_HASH_CODE

An arbitrary nonzero constant for the hash code of null valued tokens.

See Also:

Constant Field Values

Constructor Detail

HashUtil

public HashUtil()

Method Detail

binaryHashCode

public static final int binaryHashCode(byte[] bytes)

Hash a binary value using Bob Jenkins' "One-at-a-Time Hash".

Parameters:

bytes - the value to hash

Returns:

the hash code

binaryHashCode

public static final int binaryHashCode(byte[] bytes,
                                       int offset,
                                       int len)

Hash a binary value using Bob Jenkins' "One-at-a-Time Hash".

Parameters:

bytes - the value to hash

offset - start offset

len - number of bytes to code

Returns:

the hash code

intHashCode

public static final int intHashCode(int a)

Hash a binary value using Bob Jenkins' "4-byte Integer Hashing".

Parameters:

a - the value to hash

Returns:

the hash code

charHashCode

public static final int charHashCode(char c)

Hashes a character value using the same algorithm as intHashCode(int).

Parameters:

c - the value to hash

Returns:

the hash code

booleanHashCode

public static final int booleanHashCode(boolean b)

Hashes a boolean value using the same algorithm as intHashCode(int).

Parameters:

b - the value to hash

Returns:

the hash code

doubleHashCode

public static final int doubleHashCode(double d)

Hashes a double value using the same algorithm as longHashCode(long).

Parameters:

d - the value to hash

Returns:

the hash code

floatHashCode

public static final int floatHashCode(float f)

Hashes a float value using the same algorithm as intHashCode(int).

Parameters:

f - the value to hash

Returns:

the hash code

longHashCode

public static final int longHashCode(long l)

Hashes a long value.

Parameters:

l - the value to hash

Returns:

the hash code

objectHashCode

public static final int objectHashCode(Object o)

Hashes an object using Object.hashCode(). Null values are given their own hash code.

Parameters:

o - the value to hash

Returns:

the hash code

timestampHashCode

public static final int timestampHashCode(long epochSecs,
                                          int nanos)

Hashes a timestamp decomposed in seconds and nanos parts. Note that time zone offset is deliberately not included. Equal timestamp values in different time zones must have the same hashCode.

Parameters:

epochSecs - the seconds-since-epoch portion of the timestamp

nanos - the nanoseconds portion of the timestamp

Returns:

the hash code

finalMix

public static final int finalMix(int a,
                                 int b,
                                 int c)

Replicates the computation of Bob Jenkins' "final" macro.

Parameters:

a - the first input to the macro

b - the second input to the macro

c - the third input to the macro

Returns:

the resulting mix