org.apache.hadoop.hbase
Class KeyValue.MetaKeyComparator
java.lang.Object
org.apache.hadoop.hbase.KeyValue.KeyComparator
org.apache.hadoop.hbase.KeyValue.MetaKeyComparator
- All Implemented Interfaces:
- Comparator<byte[]>, KeyValue.SamePrefixComparator<byte[]>, org.apache.hadoop.io.RawComparator<byte[]>
- Enclosing class:
- KeyValue
public static class KeyValue.MetaKeyComparator
- extends KeyValue.KeyComparator
Compare key portion of a KeyValue
for keys in .META.
table.
Method Summary |
protected int |
compareRowid(byte[] left,
int loffset,
int llength,
byte[] right,
int roffset,
int rlength)
|
int |
compareRows(byte[] left,
int loffset,
int llength,
byte[] right,
int roffset,
int rlength)
|
byte[] |
getShortMidpointKey(byte[] leftKey,
byte[] rightKey)
Generate a faked byte array if possible. |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
KeyValue.MetaKeyComparator
public KeyValue.MetaKeyComparator()
compareRows
public int compareRows(byte[] left,
int loffset,
int llength,
byte[] right,
int roffset,
int rlength)
- Overrides:
compareRows
in class KeyValue.KeyComparator
getShortMidpointKey
public byte[] getShortMidpointKey(byte[] leftKey,
byte[] rightKey)
- Description copied from class:
KeyValue.KeyComparator
- Generate a faked byte array if possible. It aims to:
1)reduce key length, which expects to reduce HFile index memory footprint
2)replace TS field with LATEST_TIMESTAMP(to avoid seeking previous block)
see HBASE-7845 for more details
we need to ensure: leftKey < newKey <= rightKey
- Overrides:
getShortMidpointKey
in class KeyValue.KeyComparator
- Parameters:
leftKey
- the previous block's real stop key usuallyrightKey
- the current block's real start key usually
- Returns:
- newKey: the newly generated faked key
compareRowid
protected int compareRowid(byte[] left,
int loffset,
int llength,
byte[] right,
int roffset,
int rlength)
Copyright © 2013 The Apache Software Foundation. All Rights Reserved.