org.apache.hadoop.hive.ql.io.orc

Interface Reader

All Known Implementing Classes:

ReaderImpl

public interface Reader

The interface for reading ORC files. One Reader can support multiple concurrent RecordReader.

Nested Class Summary

Nested Classes

Modifier and Type	Interface and Description
static class	Reader.Options Options for creating a RecordReader.

Method Summary

Methods

Modifier and Type	Method and Description
CompressionKind	getCompression() Get the compression kind.
int	getCompressionSize() Get the buffer size for the compression.
long	getContentLength() Get the length of the file.
OrcFile.Version	getFileVersion() Get the file format version.
Metadata	getMetadata() Get the metadata information like stripe level column statistics etc.
List<String>	getMetadataKeys() Get the user metadata keys.
ByteBuffer	getMetadataValue(String key) Get a user metadata value.
long	getNumberOfRows() Get the number of rows in the file.
ObjectInspector	getObjectInspector() Get the object inspector for looking at the objects.
long	getRawDataSize() Get the deserialized data size of the file
long	getRawDataSizeOfColumns(List<String> colNames) Get the deserialized data size of the specified columns
int	getRowIndexStride() Get the number of rows per a entry in the row index.
ColumnStatistics[]	getStatistics() Get the statistics about the columns in the file.
List<StripeInformation>	getStripes() Get the list of stripes.
List<OrcProto.Type>	getTypes() Get the list of types contained in the file.
OrcFile.WriterVersion	getWriterVersion() Get the version of the writer of this file.
boolean	hasMetadataValue(String key) Did the user set the given metadata value.
MetadataReader	metadata()
RecordReader	rows() Create a RecordReader that reads everything with the default options.
RecordReader	rows(boolean[] include) Create a RecordReader that will scan the entire file.
RecordReader	rows(long offset, long length, boolean[] include) Create a RecordReader that will start reading at the first stripe after offset up to the stripe that starts at offset + length.
RecordReader	rows(long offset, long length, boolean[] include, SearchArgument sarg, String[] neededColumns) Create a RecordReader that will read a section of a file.
RecordReader	rowsOptions(Reader.Options options) Create a RecordReader that uses the options given.

Method Detail

getNumberOfRows

long getNumberOfRows()

Get the number of rows in the file.

Returns:

the number of rows

getRawDataSize

long getRawDataSize()

Get the deserialized data size of the file

Returns:

raw data size

getRawDataSizeOfColumns

long getRawDataSizeOfColumns(List<String> colNames)

Get the deserialized data size of the specified columns

Parameters:

colNames -

Returns:

raw data size of columns

getMetadataKeys

List<String> getMetadataKeys()

Get the user metadata keys.

Returns:

the set of metadata keys

getMetadataValue

ByteBuffer getMetadataValue(String key)

Get a user metadata value.

Parameters:

key - a key given by the user

Returns:

the bytes associated with the given key

hasMetadataValue

boolean hasMetadataValue(String key)

Did the user set the given metadata value.

Parameters:

key - the key to check

Returns:

true if the metadata value was set

getCompression

CompressionKind getCompression()

Get the compression kind.

Returns:

the kind of compression in the file

getCompressionSize

int getCompressionSize()

Get the buffer size for the compression.

Returns:

number of bytes to buffer for the compression codec.

getRowIndexStride

int getRowIndexStride()

Get the number of rows per a entry in the row index.

Returns:

the number of rows per an entry in the row index or 0 if there is no row index.

getStripes

List<StripeInformation> getStripes()

Get the list of stripes.

Returns:

the information about the stripes in order

getObjectInspector

ObjectInspector getObjectInspector()

Get the object inspector for looking at the objects.

Returns:

an object inspector for each row returned

getContentLength

long getContentLength()

Get the length of the file.

Returns:

the number of bytes in the file

getStatistics

ColumnStatistics[] getStatistics()

Get the statistics about the columns in the file.

Returns:

the information about the column

getMetadata

Metadata getMetadata()
                     throws IOException

Get the metadata information like stripe level column statistics etc.

Returns:

the information about the column

Throws:

IOException

getTypes

List<OrcProto.Type> getTypes()

Get the list of types contained in the file. The root type is the first type in the list.

Returns:

the list of flattened types

getFileVersion

OrcFile.Version getFileVersion()

Get the file format version.

getWriterVersion

OrcFile.WriterVersion getWriterVersion()

Get the version of the writer of this file.

rows

RecordReader rows()
                  throws IOException

Create a RecordReader that reads everything with the default options.

Returns:

a new RecordReader

Throws:

IOException

rowsOptions

RecordReader rowsOptions(Reader.Options options)
                         throws IOException

Create a RecordReader that uses the options given. This method can't be named rows, because many callers used rows(null) before the rows() method was introduced.

Parameters:

options - the options to read with

Returns:

a new RecordReader

Throws:

IOException

rows

RecordReader rows(boolean[] include)
                  throws IOException

Create a RecordReader that will scan the entire file. This is a legacy method and rowsOptions is preferred.

Parameters:

include - true for each column that should be included

Returns:

A new RecordReader

Throws:

IOException

rows

RecordReader rows(long offset,
                long length,
                boolean[] include)
                  throws IOException

Create a RecordReader that will start reading at the first stripe after offset up to the stripe that starts at offset + length. This is intended to work with MapReduce's FileInputFormat where divisions are picked blindly, but they must cover all of the rows. This is a legacy method and rowsOptions is preferred.

Parameters:

offset - a byte offset in the file

length - a number of bytes in the file

include - true for each column that should be included

Returns:

a new RecordReader that will read the specified rows.

Throws:

IOException

rows

RecordReader rows(long offset,
                long length,
                boolean[] include,
                SearchArgument sarg,
                String[] neededColumns)
                  throws IOException

Create a RecordReader that will read a section of a file. It starts reading at the first stripe after the offset and continues to the stripe that starts at offset + length. It also accepts a list of columns to read and a search argument. This is a legacy method and rowsOptions is preferred.

Parameters:

offset - the minimum offset of the first stripe to read

length - the distance from offset of the first address to stop reading at

include - true for each column that should be included

sarg - a search argument that limits the rows that should be read.

neededColumns - the names of the included columns

Returns:

the record reader for the rows

Throws:

IOException

metadata

MetadataReader metadata()
                        throws IOException

Throws:

IOException