org.apache.orc

Interface DataReader

All Superinterfaces:

AutoCloseable, Cloneable

public interface DataReader
extends AutoCloseable, Cloneable

An abstract data reader that IO formats can use to read bytes from underlying storage.

Method Summary

Modifier and Type	Method and Description
DataReader	clone() Clone the entire state of the DataReader with the assumption that the clone will be closed at a different time.
void	close()
CompressionCodec	getCompressionCodec() Returns the compression codec used by this datareader.
boolean	isTrackingDiskRanges() Whether the user should release buffers created by readFileData.
void	open() Opens the DataReader, making it ready to use.
DiskRangeList	readFileData(DiskRangeList range, long baseOffset, boolean doForceDirect) Reads the data.
OrcIndex	readRowIndex(StripeInformation stripe, TypeDescription fileSchema, OrcProto.StripeFooter footer, boolean ignoreNonUtf8BloomFilter, boolean[] included, OrcProto.RowIndex[] indexes, boolean[] sargColumns, OrcFile.WriterVersion version, OrcProto.Stream.Kind[] bloomFilterKinds, OrcProto.BloomFilterIndex[] bloomFilterIndices)
OrcProto.StripeFooter	readStripeFooter(StripeInformation stripe)
void	releaseBuffer(ByteBuffer toRelease) Releases buffers created by readFileData.

Method Detail

open

void open()
   throws IOException

Opens the DataReader, making it ready to use.

Throws:

IOException

readRowIndex

OrcIndex readRowIndex(StripeInformation stripe,
                      TypeDescription fileSchema,
                      OrcProto.StripeFooter footer,
                      boolean ignoreNonUtf8BloomFilter,
                      boolean[] included,
                      OrcProto.RowIndex[] indexes,
                      boolean[] sargColumns,
                      OrcFile.WriterVersion version,
                      OrcProto.Stream.Kind[] bloomFilterKinds,
                      OrcProto.BloomFilterIndex[] bloomFilterIndices)
               throws IOException

Throws:

IOException

readStripeFooter

OrcProto.StripeFooter readStripeFooter(StripeInformation stripe)
                                throws IOException

Throws:

IOException

readFileData

DiskRangeList readFileData(DiskRangeList range,
                           long baseOffset,
                           boolean doForceDirect)
                    throws IOException

Reads the data. Note that for the cases such as zero-copy read, caller must release the disk ranges produced after being done with them. Call isTrackingDiskRanges to find out if this is needed.

Parameters:

range - List if disk ranges to read. Ranges with data will be ignored.

baseOffset - Base offset from the start of the file of the ranges in disk range list.

doForceDirect - Whether the data should be read into direct buffers.

Returns:

New or modified list of DiskRange-s, where all the ranges are filled with data.

Throws:

IOException

isTrackingDiskRanges

boolean isTrackingDiskRanges()

Whether the user should release buffers created by readFileData. See readFileData javadoc.

releaseBuffer

void releaseBuffer(ByteBuffer toRelease)

Releases buffers created by readFileData. See readFileData javadoc.

Parameters:

toRelease - The buffer to release.

clone

DataReader clone()

Clone the entire state of the DataReader with the assumption that the clone will be closed at a different time. Thus, any file handles in the implementation need to be cloned.

Returns:

a new instance

close

void close()
    throws IOException

Specified by:

close in interface AutoCloseable

Throws:

IOException

getCompressionCodec

CompressionCodec getCompressionCodec()

Returns the compression codec used by this datareader. We should consider removing this from the interface.

Returns:

the compression codec