Package org.apache.druid.query.groupby.epinephelinae.collection

Class HashTableUtils

java.lang.Object

org.apache.druid.query.groupby.epinephelinae.collection.HashTableUtils

public class HashTableUtils
extends Object

Method Summary

Modifier and Type	Method	Description
static int	hashMemory(org.apache.datasketches.memory.Memory memory, long position, int length)	Compute a simple, fast hash code of some memory range.
static boolean	memoryEquals(org.apache.datasketches.memory.Memory memory1, long offset1, org.apache.datasketches.memory.Memory memory2, long offset2, int length)	Compare two memory ranges for equality.
static int	previousPowerOfTwo(int n)	Computes the previous power of two less than or equal to a given "n".

Methods inherited from class java.lang.Object

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

Method Detail

previousPowerOfTwo

public static int previousPowerOfTwo(int n)

Computes the previous power of two less than or equal to a given "n". The integer should be between 1 (inclusive) and Integer.MAX_VALUE for best results. Other parameters will return Integer.MIN_VALUE.

hashMemory

public static int hashMemory(org.apache.datasketches.memory.Memory memory,
                             long position,
                             int length)

Compute a simple, fast hash code of some memory range.

Parameters:

memory - a region of memory

position - position within the memory region

length - length of memory to hash, starting at the position

memoryEquals

public static boolean memoryEquals(org.apache.datasketches.memory.Memory memory1,
                                   long offset1,
                                   org.apache.datasketches.memory.Memory memory2,
                                   long offset2,
                                   int length)

Compare two memory ranges for equality. The purpose of this function is to be faster than BaseState.equalTo(long, java.lang.Object, long, long) for the small memory ranges that typically comprise keys in hash tables. As of this writing, it is. See "MemoryBenchmark" in the druid-benchmarks module for performance evaluation code.

Parameters:

memory1 - a region of memory

offset1 - position within the first memory region

memory2 - another region of memory

offset2 - position within the second memory region

length - length of memory to compare, starting at the positions