Package org.apache.druid.hll

Class HyperLogLogHash

java.lang.Object

org.apache.druid.hll.HyperLogLogHash

public class HyperLogLogHash
extends Object

Contains data hashing methods used before applying Hyper Log Log. HyperLogLogCollector.add(byte[]) requires hashed value on input. This class makes it easier to achieve consistent value hashing for both internal aggregation and external hashing for pre-computed aggregation.

By default 128-bit murmur3 algorithm, x64 variant is used. Implementation is thread safe.

Caution! changing of implementation may cause improper cardinality estimation between data hashed with different versions.

Constructor Summary

Constructors

Constructor	Description
HyperLogLogHash(com.google.common.hash.HashFunction hashFunction)

Method Summary

Modifier and Type	Method	Description
static HyperLogLogHash	getDefault()
byte[]	hash(byte[] rawValue)
byte[]	hash(String rawValue)

Methods inherited from class java.lang.Object

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

Constructor Detail

HyperLogLogHash

public HyperLogLogHash(com.google.common.hash.HashFunction hashFunction)

Method Detail

getDefault

public static HyperLogLogHash getDefault()

hash

public byte[] hash(byte[] rawValue)

hash

public byte[] hash(String rawValue)