com.northconcepts.datapipeline.parquet

Class ParquetDataWriter

java.lang.Object

com.northconcepts.datapipeline.core.DataObject

com.northconcepts.datapipeline.core.Endpoint

com.northconcepts.datapipeline.core.DataEndpoint

com.northconcepts.datapipeline.core.DataWriter

com.northconcepts.datapipeline.internal.lang.IntegrationWriter

com.northconcepts.datapipeline.parquet.ParquetDataWriter

public class ParquetDataWriter
extends IntegrationWriter

Writes records to Apache Parquet columnar files. See Apache Parquet columnar storage.

Nested Class Summary

Nested classes/interfaces inherited from class com.northconcepts.datapipeline.core.DataEndpoint

DataEndpoint.State

Field Summary

Fields inherited from class com.northconcepts.datapipeline.core.DataEndpoint

lastRecord, PRODUCT, PRODUCT_VERSION, VENDOR, XML_INPUT_FACTORY_KEY

Fields inherited from class com.northconcepts.datapipeline.core.Endpoint

BUFFER_SIZE, captureElapsedTime, DEFAULT_READ_BUFFER_SIZE

Fields inherited from class com.northconcepts.datapipeline.core.DataObject

id, log, name, TIMESTAMP_FORMAT

Constructor Summary

Constructors

Constructor and Description
ParquetDataWriter(File file) Write parquet data to a file.
ParquetDataWriter(OutputFile outputFile) Write parquet data to an OutputFile.

Method Summary

Modifier and Type	Method and Description
DataException	addExceptionProperties(DataException exception) Adds this endpoint's current state to a DataException.
void	close() Indicates that this endpoint has finished reading or writing.
String	getCacheFolder() Indicates the folder to store cached files during dynamic schema generation with LocalFileDataset or null if schema generation should be performed completely in memory using MemoryDataset /> See also getRecordsPerCacheFile()
protected int	getColumnStatsReaderThreads()
CompressionCodecName	getCompressionCodecName() Indicates the compression used for writing (default UNCOMPRESSED).
Configuration	getConfiguration() Returns the Parquet configuration parameters.
int	getDefaultBigDecimalScale() Returns the default scale used when writing BigDecimal values (default 5).
int	getDefaultBigNumberPrecision() Returns the default precision used when writing BigDecimal & BigInteger values (default 25).
Long	getMaxRecordsAnalyzed() Indicates how many records should be analyzed and cached to generate the Parquet schema if no schema was explicitly set on this writer (default is 1000).
int	getRecordsPerCacheFile() Indicates how many records should be cached in memory to generate the Parquet schema if no schema was explicitly set on this writer (default is 10_000L).
RoundingMode	getRoundingMode() Indicates the rounding algorithm used for all BigDecimal values (default is RoundingMode.HALF_UP).
MessageType	getSchema() Returns the schema used to write the file.
boolean	isDefaulAdjustToUTC() Deprecated. see isDefaultAdjustedToUTC()
boolean	isDefaultAdjustedToUTC() Indicates if all datetime fields should be marked as AdjustedToUTC.
void	open() Makes this endpoint ready for reading or writing.
ParquetDataWriter	setCacheFolder(String cacheFolder) Indicates the folder to store cached files during dynamic schema generation with LocalFileDataset or null if schema generation should be performed completely in memory using MemoryDataset. See also setRecordsPerCacheFile(int)
protected ParquetDataWriter	setColumnStatsReaderThreads(int columnStatsReaderThreads)
ParquetDataWriter	setCompressionCodecName(CompressionCodecName compressionCodecName) Indicates the compression used for writing (default UNCOMPRESSED).
ParquetDataWriter	setConfiguration(Configuration configuration) Sets the Parquet configuration parameters.
ParquetDataWriter	setDefaultAdjustedToUTC(boolean defaultAdjustedToUTC) Indicates if all datetime fields should be marked as AdjustedToUTC.
ParquetDataWriter	setDefaultAdjustToUTC(boolean defaultAdjustToUTC) Deprecated. see setDefaultAdjustedToUTC(boolean)
ParquetDataWriter	setDefaultBigDecimalScale(int defaultBigDecimalScale) Sets the default scale used when writing BigDecimal values (default 5).
ParquetDataWriter	setDefaultBigNumberPrecision(int defaultBigNumberPrecision) Sets the default precision used when writing BigDecimal & BigInteger values (default 25).
ParquetDataWriter	setMaxRecordsAnalyzed(Long maxRecordsAnalyzed) Indicates how many records should be analyzed and cached to generate the Parquet schema if no schema was explicitly set on this writer (default is 1000).
ParquetDataWriter	setRecordsPerCacheFile(int recordsPerCacheFile) Indicates how many records should be cached in memory to generate the Parquet schema if no schema was explicitly set on this writer (default is 10_000L).
ParquetDataWriter	setRoundingMode(RoundingMode roundingMode) Indicates the rounding algorithm used for all BigDecimal values (default is RoundingMode.HALF_UP).
ParquetDataWriter	setSchema(Connection connection, JdbcValueReader jdbcValueReader, String query, Object... queryParameters) Sets the schema used to write the file by copying it from the metadata of an SQL query.
ParquetDataWriter	setSchema(Connection connection, JdbcValueReader sqlToJavaTypeMapper, String databaseCatalog, String databaseSchema, String databaseTable) Sets the schema used to write the Parquet file by copying it from the schema of a database table.
ParquetDataWriter	setSchema(Connection connection, String query, Object... queryParameters) Sets the schema used to write the file by copying it from the metadata of an SQL query.
ParquetDataWriter	setSchema(Connection connection, String databaseCatalog, String databaseSchema, String databaseTable) Sets the schema used to write the Parquet file by copying it from the schema of a database table.
ParquetDataWriter	setSchema(JdbcConnectionFactory jdbcConnectionFactory, JdbcValueReader jdbcValueReader, String query, Object... queryParameters) Sets the schema used to write the file by copying it from the metadata of an SQL query.
ParquetDataWriter	setSchema(JdbcConnectionFactory jdbcConnectionFactory, String query, Object... queryParameters) Sets the schema used to write the file by copying it from the metadata of an SQL query.
ParquetDataWriter	setSchema(MessageType schema) Sets the schema used to write the file.
protected void	writeImpl(Record record) Overridden by subclasses to write the specified record to this DataWriter.

Methods inherited from class com.northconcepts.datapipeline.core.DataWriter

available, getNestedEndpoint, getNestedWriter, getRootEndpoint, getRootWriter, write

Methods inherited from class com.northconcepts.datapipeline.core.DataEndpoint

decrementRecordCount, enableJmx, getLastRecord, getRecordCount, getRecordCountAsBigInteger, getRecordCountAsString, incrementRecordCount, isRecordCountBigInteger, resetRecordCount, toString

Methods inherited from class com.northconcepts.datapipeline.core.Endpoint

addElapsedtime, assertClosed, assertNotOpened, assertOpened, finalize, getClosedOn, getDescription, getElapsedTime, getElapsedTimeAsString, getOpenedOn, getOpenElapsedTime, getOpenElapsedTimeAsString, getSelfTime, getSelfTimeAsString, getState, isCaptureElapsedTime, isClosed, isOpen, setCaptureElapsedTime, setDescription

Methods inherited from class com.northconcepts.datapipeline.core.DataObject

exception, exception, exception, getId, getName, resetID

Methods inherited from class java.lang.Object

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

Constructor Detail

ParquetDataWriter

public ParquetDataWriter(File file)

Write parquet data to a file.

ParquetDataWriter

public ParquetDataWriter(OutputFile outputFile)

Write parquet data to an OutputFile.

Parameters:

outputFile - - OutputFile with FileSystem.

Method Detail

open

public void open()
          throws DataException

Description copied from class: DataEndpoint

Makes this endpoint ready for reading or writing.

Overrides:

open in class IntegrationWriter

Throws:

DataException

writeImpl

protected void writeImpl(Record record)
                  throws Throwable

Description copied from class: DataWriter

Overridden by subclasses to write the specified record to this DataWriter.

Specified by:

writeImpl in class DataWriter

Throws:

Throwable

close

public void close()
           throws DataException

Description copied from class: DataEndpoint

Indicates that this endpoint has finished reading or writing.

Overrides:

close in class DataEndpoint

Throws:

DataException

getSchema

public MessageType getSchema()

Returns the schema used to write the file.

setSchema

public ParquetDataWriter setSchema(MessageType schema)

Sets the schema used to write the file.

setSchema

public ParquetDataWriter setSchema(Connection connection,
                                   String query,
                                   Object... queryParameters)

Sets the schema used to write the file by copying it from the metadata of an SQL query.

setSchema

public ParquetDataWriter setSchema(Connection connection,
                                   JdbcValueReader jdbcValueReader,
                                   String query,
                                   Object... queryParameters)

Sets the schema used to write the file by copying it from the metadata of an SQL query. Since no data is read from query (only its metadata is used), care should be taken to create an optimized query where the database does minimal work and returns no data. For example, consider using a query like: SELECT * FROM invoices WHERE 1<0.

setSchema

public ParquetDataWriter setSchema(JdbcConnectionFactory jdbcConnectionFactory,
                                   String query,
                                   Object... queryParameters)

Sets the schema used to write the file by copying it from the metadata of an SQL query. Since no data is read from query (only its metadata is used), care should be taken to create an optimized query where the database does minimal work and returns no data. For example, consider using a query like: SELECT * FROM invoices WHERE 1<0.

setSchema

public ParquetDataWriter setSchema(JdbcConnectionFactory jdbcConnectionFactory,
                                   JdbcValueReader jdbcValueReader,
                                   String query,
                                   Object... queryParameters)

Sets the schema used to write the file by copying it from the metadata of an SQL query.

setSchema

public ParquetDataWriter setSchema(Connection connection,
                                   String databaseCatalog,
                                   String databaseSchema,
                                   String databaseTable)

Sets the schema used to write the Parquet file by copying it from the schema of a database table.

setSchema

public ParquetDataWriter setSchema(Connection connection,
                                   JdbcValueReader sqlToJavaTypeMapper,
                                   String databaseCatalog,
                                   String databaseSchema,
                                   String databaseTable)

Sets the schema used to write the Parquet file by copying it from the schema of a database table.

getDefaultBigDecimalScale

public int getDefaultBigDecimalScale()

Returns the default scale used when writing BigDecimal values (default 5).

setDefaultBigDecimalScale

public ParquetDataWriter setDefaultBigDecimalScale(int defaultBigDecimalScale)

Sets the default scale used when writing BigDecimal values (default 5).

getDefaultBigNumberPrecision

public int getDefaultBigNumberPrecision()

Returns the default precision used when writing BigDecimal & BigInteger values (default 25).

setDefaultBigNumberPrecision

public ParquetDataWriter setDefaultBigNumberPrecision(int defaultBigNumberPrecision)

Sets the default precision used when writing BigDecimal & BigInteger values (default 25).

addExceptionProperties

public DataException addExceptionProperties(DataException exception)

Description copied from class: Endpoint

Adds this endpoint's current state to a DataException. Since this method is called whenever an exception is thrown, subclasses should override it to add their specific information.

Overrides:

addExceptionProperties in class DataWriter

getCompressionCodecName

public CompressionCodecName getCompressionCodecName()

Indicates the compression used for writing (default UNCOMPRESSED).

setCompressionCodecName

public ParquetDataWriter setCompressionCodecName(CompressionCodecName compressionCodecName)

Indicates the compression used for writing (default UNCOMPRESSED).

isDefaulAdjustToUTC

@Deprecated
public boolean isDefaulAdjustToUTC()

Deprecated. see isDefaultAdjustedToUTC()

setDefaultAdjustToUTC

@Deprecated
public ParquetDataWriter setDefaultAdjustToUTC(boolean defaultAdjustToUTC)

Deprecated. see setDefaultAdjustedToUTC(boolean)

isDefaultAdjustedToUTC

public boolean isDefaultAdjustedToUTC()

Indicates if all datetime fields should be marked as AdjustedToUTC.

setDefaultAdjustedToUTC

public ParquetDataWriter setDefaultAdjustedToUTC(boolean defaultAdjustedToUTC)

Indicates if all datetime fields should be marked as AdjustedToUTC.

getRoundingMode

public RoundingMode getRoundingMode()

Indicates the rounding algorithm used for all BigDecimal values (default is RoundingMode.HALF_UP).

setRoundingMode

public ParquetDataWriter setRoundingMode(RoundingMode roundingMode)

Indicates the rounding algorithm used for all BigDecimal values (default is RoundingMode.HALF_UP).

getConfiguration

public Configuration getConfiguration()

Returns the Parquet configuration parameters.

setConfiguration

public ParquetDataWriter setConfiguration(Configuration configuration)

Sets the Parquet configuration parameters.

getColumnStatsReaderThreads

protected int getColumnStatsReaderThreads()

setColumnStatsReaderThreads

protected ParquetDataWriter setColumnStatsReaderThreads(int columnStatsReaderThreads)

getMaxRecordsAnalyzed

public Long getMaxRecordsAnalyzed()

Indicates how many records should be analyzed and cached to generate the Parquet schema if no schema was explicitly set on this writer (default is 1000). This value will not be used if a schema was set on this writer.

Passing in null will cause all records to be read and cached to determine the schema.

The value will be set to 1 if a value less than 1 is passed in.

Note: Using null or a high record count can significantly slow down processing and cause an OutOfMemoryError.

setMaxRecordsAnalyzed

public ParquetDataWriter setMaxRecordsAnalyzed(Long maxRecordsAnalyzed)

Indicates how many records should be analyzed and cached to generate the Parquet schema if no schema was explicitly set on this writer (default is 1000). This value will not be used if a schema was set on this writer.

Passing in null will cause all records to be read and cached to determine the schema.

The value will be set to 1 if a value less than 1 is passed in.

Note: Using null or a high record count can significantly slow down processing and cause an OutOfMemoryError.

getRecordsPerCacheFile

public int getRecordsPerCacheFile()

Indicates how many records should be cached in memory to generate the Parquet schema if no schema was explicitly set on this writer (default is 10_000L). This value will not be used if a schema was set on this writer.

The value will be set to 10_000L if a value less than 1 is passed in.

Note: Using a small record count can significantly slow down processing due to many IO operations. And using a very huge record count can reduce the performance and consume more memory as it will hold more records in memory.

Used only if setCacheFolder(String) is set.

setRecordsPerCacheFile

public ParquetDataWriter setRecordsPerCacheFile(int recordsPerCacheFile)

Indicates how many records should be cached in memory to generate the Parquet schema if no schema was explicitly set on this writer (default is 10_000L). This value will not be used if a schema was set on this writer.

The value will be set to 10_000 if a value less than 1 is passed in.

Note: Using a small record count can significantly slow down processing due to many IO operations. And using a very huge record count can reduce the performance and consume more memory as it will hold more records in memory.

Used only if setCacheFolder(String) is set.

getCacheFolder

public String getCacheFolder()

Indicates the folder to store cached files during dynamic schema generation with LocalFileDataset or null if schema generation should be performed completely in memory using MemoryDataset /> See also getRecordsPerCacheFile()

setCacheFolder

public ParquetDataWriter setCacheFolder(String cacheFolder)

Indicates the folder to store cached files during dynamic schema generation with LocalFileDataset or null if schema generation should be performed completely in memory using MemoryDataset.
See also setRecordsPerCacheFile(int)