Package org.apache.hadoop.hbase.io.hfile

Class HFileBlockIndex

java.lang.Object

org.apache.hadoop.hbase.io.hfile.HFileBlockIndex

@Private
public class HFileBlockIndex
extends Object

Provides functionality to write (HFileBlockIndex.BlockIndexWriter) and read BlockIndexReader single-level and multi-level block indexes. Examples of how to use the block index writer can be found in CompoundBloomFilterWriter and HFileWriterImpl. Examples of how to use the reader can be found in HFileReaderImpl and org.apache.hadoop.hbase.io.hfile.TestHFileBlockIndex.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
(package private) static class	HFileBlockIndex.BlockIndexChunkImpl	A single chunk of the block index in the process of writing.
(package private) static class	HFileBlockIndex.BlockIndexReader	The reader will always hold the root level index in the memory.
static class	HFileBlockIndex.BlockIndexWriter	Writes the block index into the output stream.
(package private) static class	HFileBlockIndex.ByteArrayKeyBlockIndexReader	An implementation of the BlockIndexReader that deals with block keys which are plain byte[] like MetaBlock or the Bloom Block for ROW bloom.
(package private) static class	HFileBlockIndex.CellBasedKeyBlockIndexReader	An implementation of the BlockIndexReader that deals with block keys which are the key part of a cell like the Data block index or the ROW_COL bloom blocks This needs a comparator to work with the Cells
(package private) static class	HFileBlockIndex.CellBasedKeyBlockIndexReaderV2

Field Summary

Fields

Modifier and Type	Field	Description
(package private) static final int	DEFAULT_MAX_CHUNK_SIZE
(package private) static final int	DEFAULT_MIN_INDEX_NUM_ENTRIES
private static final String	INLINE_BLOCKS_NOT_ALLOWED	Error message when trying to use inline block API in single-level mode.
private static final org.slf4j.Logger	LOG
static final String	MAX_CHUNK_SIZE_KEY	The maximum size guideline for index blocks (both leaf, intermediate, and root).
protected static final int	MID_KEY_METADATA_SIZE	The size of a meta-data record used for finding the mid-key in a multi-level index.
static final String	MIN_INDEX_NUM_ENTRIES_KEY	Minimum number of entries in a single index block.
(package private) static final int	SECONDARY_INDEX_ENTRY_OVERHEAD	The number of bytes stored in each "secondary index" entry in addition to key bytes in the non-root index block format.

Constructor Summary

Constructors

Constructor	Description
HFileBlockIndex()

Method Summary

Modifier and Type	Method	Description
static int	getMaxChunkSize(org.apache.hadoop.conf.Configuration conf)
static int	getMinIndexNumEntries(org.apache.hadoop.conf.Configuration conf)

Methods inherited from class java.lang.Object

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

Field Details

LOG

private static final org.slf4j.Logger LOG

DEFAULT_MAX_CHUNK_SIZE

static final int DEFAULT_MAX_CHUNK_SIZE

See Also:

• Constant Field Values

MAX_CHUNK_SIZE_KEY

public static final String MAX_CHUNK_SIZE_KEY

The maximum size guideline for index blocks (both leaf, intermediate, and root). If not specified, DEFAULT_MAX_CHUNK_SIZE is used.

See Also:

• Constant Field Values

MIN_INDEX_NUM_ENTRIES_KEY

public static final String MIN_INDEX_NUM_ENTRIES_KEY

Minimum number of entries in a single index block. Even if we are above the hfile.index.block.max.size we will keep writing to the same block unless we have that many entries. We should have at least a few entries so that we don't have too many levels in the multi-level index. This should be at least 2 to make sure there is no infinite recursion.

See Also:

• Constant Field Values

DEFAULT_MIN_INDEX_NUM_ENTRIES

static final int DEFAULT_MIN_INDEX_NUM_ENTRIES

See Also:

• Constant Field Values

SECONDARY_INDEX_ENTRY_OVERHEAD

static final int SECONDARY_INDEX_ENTRY_OVERHEAD

The number of bytes stored in each "secondary index" entry in addition to key bytes in the non-root index block format. The first long is the file offset of the deeper-level block the entry points to, and the int that follows is that block's on-disk size without including header.

See Also:

• Constant Field Values

INLINE_BLOCKS_NOT_ALLOWED

private static final String INLINE_BLOCKS_NOT_ALLOWED

Error message when trying to use inline block API in single-level mode.

See Also:

• Constant Field Values

MID_KEY_METADATA_SIZE

protected static final int MID_KEY_METADATA_SIZE

The size of a meta-data record used for finding the mid-key in a multi-level index. Consists of the middle leaf-level index block offset (long), its on-disk size without header included (int), and the mid-key entry's zero-based index in that leaf index block.

See Also:

• Constant Field Values

Constructor Details

HFileBlockIndex

public HFileBlockIndex()

Method Details

getMaxChunkSize

public static int getMaxChunkSize(org.apache.hadoop.conf.Configuration conf)

getMinIndexNumEntries

public static int getMinIndexNumEntries(org.apache.hadoop.conf.Configuration conf)