Package org.apache.datasketches.memory

Class MurmurHash3

java.lang.Object
org.apache.datasketches.memory.MurmurHash3
public final class MurmurHash3 extends Object

The MurmurHash3 is a fast, non-cryptographic, 128-bit hash function that has excellent avalanche and 2-way bit independence properties.

Austin Appleby's C++ MurmurHash3_x64_128(...), final revision 150, which is in the Public Domain, was the inspiration for this implementation in Java.

This implementation of the MurmurHash3 allows hashing of a block of on-heap Memory defined by an offset and length. The calling API also allows the user to supply the small output array of two longs, so that the entire hash function is static and free of object allocations.

This implementation produces exactly the same hash result as the MurmurHash3 function in datasketches-java given compatible inputs.

This version 3 of the implementation leverages the jdk.incubator.foreign package of JDK-17 in place of the Unsafe class.

Author:
Lee Rhodes

  • Constructor Summary

    Constructors
    Constructor
    Description
    MurmurHash3()
     

  • Method Summary

    Modifier and Type
    Method
    Description
    static long[]
    hash(byte[] in, long seed)
    Returns a 128-bit hash of the input.
    static long[]
    hash(char[] in, long seed)
    Returns a 128-bit hash of the input.
    static long[]
    hash(double in, long seed, long[] hashOut)
    Returns a 128-bit hash of the input.
    static long[]
    hash(int[] in, long seed)
    Returns a 128-bit hash of the input.
    static long[]
    hash(long[] in, long seed)
    Returns a 128-bit hash of the input.
    static long[]
    hash(long in, long seed, long[] hashOut)
    Returns a 128-bit hash of the input.
    static long[]
    hash(String in, long seed, long[] hashOut)
    Returns a 128-bit hash of the input.
    static long[]
    hash(MemorySegment seg, long offsetBytes, long lengthBytes, long seed, long[] hashOut)
    Returns a 128-bit hash of the input as a long array of size 2.
    static long[]
    hash(Memory mem, long offsetBytes, long lengthBytes, long seed, long[] hashOut)
    Returns a 128-bit hash of the input as a long array of size 2.

    Methods inherited from class java.lang.Object

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

  • Constructor Details

    • MurmurHash3

      public MurmurHash3()

  • Method Details

    • hash

      public static long[] hash(long[] in, long seed)
      Returns a 128-bit hash of the input. Provided for compatibility with older version of MurmurHash3, but empty or null input now throws IllegalArgumentException.
      Parameters:
      in - long array
      seed - A long valued seed.
      Returns:
      the hash

    • hash

      public static long[] hash(int[] in, long seed)
      Returns a 128-bit hash of the input. Provided for compatibility with older version of MurmurHash3, but empty or null input now throws IllegalArgumentException.
      Parameters:
      in - int array
      seed - A long valued seed.
      Returns:
      the hash

    • hash

      public static long[] hash(char[] in, long seed)
      Returns a 128-bit hash of the input. Provided for compatibility with older version of MurmurHash3, but empty or null input now throws IllegalArgumentException.
      Parameters:
      in - char array
      seed - A long valued seed.
      Returns:
      the hash

    • hash

      public static long[] hash(byte[] in, long seed)
      Returns a 128-bit hash of the input. Provided for compatibility with older version of MurmurHash3, but empty or null input now throws IllegalArgumentException.
      Parameters:
      in - byte array
      seed - A long valued seed.
      Returns:
      the hash

    • hash

      public static long[] hash(long in, long seed, long[] hashOut)
      Returns a 128-bit hash of the input. Note the entropy of the resulting hash cannot be more than 64 bits.
      Parameters:
      in - a long
      seed - A long valued seed.
      hashOut - A long array of size 2
      Returns:
      the hash

    • hash

      public static long[] hash(double in, long seed, long[] hashOut)
      Returns a 128-bit hash of the input. Note the entropy of the resulting hash cannot be more than 64 bits.
      Parameters:
      in - a double
      seed - A long valued seed.
      hashOut - A long array of size 2
      Returns:
      the hash

    • hash

      public static long[] hash(String in, long seed, long[] hashOut)
      Returns a 128-bit hash of the input. An empty or null input throws IllegalArgumentException.
      Parameters:
      in - a String
      seed - A long valued seed.
      hashOut - A long array of size 2
      Returns:
      the hash

    • hash

      public static long[] hash(Memory mem, long offsetBytes, long lengthBytes, long seed, long[] hashOut)
      Returns a 128-bit hash of the input as a long array of size 2.
      Parameters:
      mem - The input on-heap Memory. Must be non-null and non-empty.
      offsetBytes - the starting point within Memory.
      lengthBytes - the total number of bytes to be hashed.
      seed - A long valued seed.
      hashOut - the size 2 long array for the resulting 128-bit hash
      Returns:
      the hash.

    • hash

      public static long[] hash(MemorySegment seg, long offsetBytes, long lengthBytes, long seed, long[] hashOut)
      Returns a 128-bit hash of the input as a long array of size 2.
      Parameters:
      seg - The input MemorySegment. Must be non-null and non-empty.
      offsetBytes - the starting point within Memory.
      lengthBytes - the total number of bytes to be hashed.
      seed - A long valued seed.
      hashOut - the size 2 long array for the resulting 128-bit hash
      Returns:
      the hash.