org.apache.hadoop.hbase.io.compress

Class CompressionUtil

java.lang.Object

org.apache.hadoop.hbase.io.compress.CompressionUtil

@InterfaceAudience.Private
public final class CompressionUtil
extends Object

Constructor Summary

Constructors

Modifier	Constructor and Description
private	CompressionUtil()

Method Summary

Modifier and Type	Method and Description
static int	compressionOverhead(int bufferSize) Most compression algorithms can be presented with pathological input that causes an expansion rather than a compression.
static int	roundInt2(int v) Round up to the next power of two, unless the value would become negative (ints are signed), in which case just return Integer.MAX_VALUE.

Methods inherited from class java.lang.Object

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

Constructor Detail

CompressionUtil

private CompressionUtil()

Method Detail

roundInt2

public static int roundInt2(int v)

Round up to the next power of two, unless the value would become negative (ints are signed), in which case just return Integer.MAX_VALUE.

compressionOverhead

public static int compressionOverhead(int bufferSize)

Most compression algorithms can be presented with pathological input that causes an expansion rather than a compression. Hadoop's compression API requires that we calculate additional buffer space required for the worst case. There is a formula developed for gzip that applies as a ballpark to all LZ variants. It should be good enough for now and has been tested as such with a range of different inputs.