org.apache.hadoop.hbase.io.hfile

Interface HFileDataBlockEncoder

All Known Implementing Classes:

HFileDataBlockEncoderImpl, NoOpDataBlockEncoder

@InterfaceAudience.Private
public interface HFileDataBlockEncoder

Controls what kind of data block encoding is used. If data block encoding is not set or the given block is not a data block (encoded or not), methods should just return the unmodified block.

Field Summary

Fields

Modifier and Type	Field and Description
static byte[]	DATA_BLOCK_ENCODING Type of encoding used for data blocks in HFile.

Method Summary

Modifier and Type	Method and Description
void	encode(Cell cell, HFileBlockEncodingContext encodingCtx, DataOutputStream out) Encodes a KeyValue.
void	endBlockEncoding(HFileBlockEncodingContext encodingCtx, DataOutputStream out, byte[] uncompressedBytesWithHeader, BlockType blockType) Ends encoding for a block of KeyValues.
DataBlockEncoding	getDataBlockEncoding() Returns the data block encoding
DataBlockEncoding	getEffectiveEncodingInCache(boolean isCompaction)
HFileBlockDecodingContext	newDataBlockDecodingContext(org.apache.hadoop.conf.Configuration conf, HFileContext fileContext) create a encoder specific decoding context for reading.
HFileBlockEncodingContext	newDataBlockEncodingContext(org.apache.hadoop.conf.Configuration conf, byte[] headerBytes, HFileContext fileContext) Create an encoder specific encoding context object for writing.
void	saveMetadata(HFile.Writer writer) Save metadata in HFile which will be written to disk
void	startBlockEncoding(HFileBlockEncodingContext encodingCtx, DataOutputStream out) Starts encoding for a block of KeyValues.
boolean	useEncodedScanner() Decides whether we should use a scanner over encoded blocks.

Field Detail

DATA_BLOCK_ENCODING

static final byte[] DATA_BLOCK_ENCODING

Type of encoding used for data blocks in HFile. Stored in file info.

Method Detail

startBlockEncoding

void startBlockEncoding(HFileBlockEncodingContext encodingCtx,
                        DataOutputStream out)
                 throws IOException

Starts encoding for a block of KeyValues. Call endBlockEncoding(HFileBlockEncodingContext, DataOutputStream, byte[], BlockType) to finish encoding of a block. nnn

Throws:

IOException

encode

void encode(Cell cell,
            HFileBlockEncodingContext encodingCtx,
            DataOutputStream out)
     throws IOException

Encodes a KeyValue. nnnn

Throws:

IOException

endBlockEncoding

void endBlockEncoding(HFileBlockEncodingContext encodingCtx,
                      DataOutputStream out,
                      byte[] uncompressedBytesWithHeader,
                      BlockType blockType)
               throws IOException

Ends encoding for a block of KeyValues. Gives a chance for the encoder to do the finishing stuff for the encoded block. It must be called at the end of block encoding. nnnnn

Throws:

IOException

useEncodedScanner

boolean useEncodedScanner()

Decides whether we should use a scanner over encoded blocks.

Returns:

Whether to use encoded scanner.

saveMetadata

void saveMetadata(HFile.Writer writer)
           throws IOException

Save metadata in HFile which will be written to disk

Parameters:

writer - writer for a given HFile

Throws:

IOException - on disk problems

getDataBlockEncoding

DataBlockEncoding getDataBlockEncoding()

Returns the data block encoding

getEffectiveEncodingInCache

DataBlockEncoding getEffectiveEncodingInCache(boolean isCompaction)

Returns:

the effective in-cache data block encoding, taking into account whether we are doing a compaction.

newDataBlockEncodingContext

HFileBlockEncodingContext newDataBlockEncodingContext(org.apache.hadoop.conf.Configuration conf,
                                                      byte[] headerBytes,
                                                      HFileContext fileContext)

Create an encoder specific encoding context object for writing. And the encoding context should also perform compression if compressionAlgorithm is valid.

Parameters:

conf - store configuration

headerBytes - header bytes

fileContext - HFile meta data

Returns:

a new HFileBlockEncodingContext object

newDataBlockDecodingContext

HFileBlockDecodingContext newDataBlockDecodingContext(org.apache.hadoop.conf.Configuration conf,
                                                      HFileContext fileContext)

create a encoder specific decoding context for reading. And the decoding context should also do decompression if compressionAlgorithm is valid.

Parameters:

conf - store configuration

fileContext - - HFile meta data

Returns:

a new HFileBlockDecodingContext object