org.apache.hadoop.hbase.codec.prefixtree.encode.row

Class RowNodeWriter

java.lang.Object

org.apache.hadoop.hbase.codec.prefixtree.encode.row.RowNodeWriter

@InterfaceAudience.Private
public class RowNodeWriter
extends Object

Serializes the fields comprising one node of the row trie, which can be a branch, nub, or leaf. Please see the write() method for the order in which data is written.

Field Summary

Fields

Modifier and Type	Field and Description
protected PrefixTreeBlockMeta	blockMeta
protected int	fanOut
protected static Log	LOG
protected int	numCells
protected PrefixTreeEncoder	prefixTreeEncoder fields
protected TokenizerNode	tokenizerNode
protected int	tokenWidth
protected int	width

Constructor Summary

Constructors

Constructor and Description
RowNodeWriter(PrefixTreeEncoder keyValueBuilder, TokenizerNode tokenizerNode) construct

Method Summary

Modifier and Type	Method and Description
protected void	calculateOffsetsAndLengths() methods
int	calculateWidth()
int	calculateWidthOverrideOffsetWidth(int offsetWidth)
void	reconstruct(PrefixTreeEncoder prefixTreeEncoder, TokenizerNode tokenizerNode)
void	reset(TokenizerNode node)
void	write(OutputStream os) writing the compiled structure to the OutputStream
protected void	writeCellTypes(OutputStream os)
protected void	writeFamilyNodeOffsets(OutputStream os) The following methods write data for each cell in the row, mostly consisting of indexes or offsets into the timestamp/column data structures that are written in the middle of the block.
void	writeFan(OutputStream os) UVInt: numFanBytes/fanOut bytes: each fan byte
protected void	writeMvccVersionIndexes(OutputStream os)
protected void	writeNextRowTrieNodeOffsets(OutputStream os) If a branch or a nub, the last thing we append are the UFInt offsets to the child row nodes.
protected void	writeNumCells(OutputStream os) UVInt: numCells, the number of cells in this row which will be 0 for branch nodes
protected void	writeQualifierNodeOffsets(OutputStream os)
protected void	writeRowToken(OutputStream os) UVInt: tokenWidth bytes: token
protected void	writeTagNodeOffsets(OutputStream os)
protected void	writeTimestampIndexes(OutputStream os)
protected void	writeValueLengths(OutputStream os)
protected void	writeValueOffsets(OutputStream os)

Methods inherited from class java.lang.Object

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

Field Detail

LOG

protected static final Log LOG

prefixTreeEncoder

protected PrefixTreeEncoder prefixTreeEncoder

fields

blockMeta

protected PrefixTreeBlockMeta blockMeta

tokenizerNode

protected TokenizerNode tokenizerNode

tokenWidth

protected int tokenWidth

fanOut

protected int fanOut

numCells

protected int numCells

width

protected int width

Constructor Detail

RowNodeWriter

public RowNodeWriter(PrefixTreeEncoder keyValueBuilder,
                     TokenizerNode tokenizerNode)

construct

Method Detail

reconstruct

public void reconstruct(PrefixTreeEncoder prefixTreeEncoder,
                        TokenizerNode tokenizerNode)

reset

public void reset(TokenizerNode node)

calculateOffsetsAndLengths

protected void calculateOffsetsAndLengths()

methods

calculateWidth

public int calculateWidth()

calculateWidthOverrideOffsetWidth

public int calculateWidthOverrideOffsetWidth(int offsetWidth)

write

public void write(OutputStream os)
           throws IOException

writing the compiled structure to the OutputStream

Throws:

IOException

writeRowToken

protected void writeRowToken(OutputStream os)
                      throws IOException

UVInt: tokenWidth bytes: token

Throws:

IOException

writeFan

public void writeFan(OutputStream os)
              throws IOException

UVInt: numFanBytes/fanOut bytes: each fan byte

Throws:

IOException

writeNumCells

protected void writeNumCells(OutputStream os)
                      throws IOException

UVInt: numCells, the number of cells in this row which will be 0 for branch nodes

Throws:

IOException

writeFamilyNodeOffsets

protected void writeFamilyNodeOffsets(OutputStream os)
                               throws IOException

The following methods write data for each cell in the row, mostly consisting of indexes or offsets into the timestamp/column data structures that are written in the middle of the block. We use UFIntTool to encode these indexes/offsets to allow random access during a binary search of a particular column/timestamp combination.

Branch nodes will not have any data in these sections.

Throws:

IOException

writeQualifierNodeOffsets

protected void writeQualifierNodeOffsets(OutputStream os)
                                  throws IOException

Throws:

IOException

writeTagNodeOffsets

protected void writeTagNodeOffsets(OutputStream os)
                            throws IOException

Throws:

IOException

writeTimestampIndexes

protected void writeTimestampIndexes(OutputStream os)
                              throws IOException

Throws:

IOException

writeMvccVersionIndexes

protected void writeMvccVersionIndexes(OutputStream os)
                                throws IOException

Throws:

IOException

writeCellTypes

protected void writeCellTypes(OutputStream os)
                       throws IOException

Throws:

IOException

writeValueOffsets

protected void writeValueOffsets(OutputStream os)
                          throws IOException

Throws:

IOException

writeValueLengths

protected void writeValueLengths(OutputStream os)
                          throws IOException

Throws:

IOException

writeNextRowTrieNodeOffsets

protected void writeNextRowTrieNodeOffsets(OutputStream os)
                                    throws IOException

If a branch or a nub, the last thing we append are the UFInt offsets to the child row nodes.

Throws:

IOException