com.actian.dataflow.hive.shims

Class ParquetFileShimImpl

java.lang.Object

com.actian.dataflow.hive.shims.ParquetFileShimImpl

All Implemented Interfaces:

ParquetFileShim

public class ParquetFileShimImpl
extends Object
implements ParquetFileShim

Shim layer implementation for Parquet files.

Constructor Summary

Constructors

Constructor and Description
ParquetFileShimImpl(FileClient fileClient, Path filePath)

Method Summary

Methods

Modifier and Type	Method and Description
Path	getFilePath() Get the file path of the Parquet file.
RecordReaderShim	getReader(long offset, long length, String[] columns, RecordSettable target) Get a reader that can read records from the Parquet file.
RecordTokenType	getType() Get the record type of the Parquet file.

Methods inherited from class java.lang.Object

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

Constructor Detail

ParquetFileShimImpl

public ParquetFileShimImpl(FileClient fileClient,
                   Path filePath)
                    throws IOException

Throws:

IOException

Method Detail

getFilePath

public Path getFilePath()

Description copied from interface: ParquetFileShim

Get the file path of the Parquet file.

Specified by:

getFilePath in interface ParquetFileShim

Returns:

file path

getType

public RecordTokenType getType()
                        throws IOException

Description copied from interface: ParquetFileShim

Get the record type of the Parquet file. Note that unsupported Parquet types will be ignored with a warning.

Specified by:

getType in interface ParquetFileShim

Returns:

record type of the Parquet file

Throws:

IOException - thrown if an I/O error occurs accessing the Parquet file

getReader

public RecordReaderShim getReader(long offset,
                         long length,
                         String[] columns,
                         RecordSettable target)
                           throws IOException

Description copied from interface: ParquetFileShim

Get a reader that can read records from the Parquet file.

Specified by:

getReader in interface ParquetFileShim

Parameters:

offset - beginning offset in the file to begin reading (start of split)

length - amount of data to read (size of split)

columns - names of the columns to read (empty = all columns)

target - record target for the data read by the reader

Returns:

a shim over the Parquet record reader

Throws:

IOException - thrown if an I/O error occurs accessing the Parquet file