Package io.trino.plugin.hive.parquet

Class ParquetPageSourceFactory

java.lang.Object

io.trino.plugin.hive.parquet.ParquetPageSourceFactory

All Implemented Interfaces:

HivePageSourceFactory

public class ParquetPageSourceFactory
extends Object
implements HivePageSourceFactory

Field Summary

Fields

Modifier and Type	Field	Description
static HiveColumnHandle	PARQUET_ROW_INDEX_COLUMN	If this object is passed as one of the columns for createPageSource, it will be populated as an additional column containing the index of each row read.

Constructor Summary

Constructors

Constructor	Description
ParquetPageSourceFactory(HdfsEnvironment hdfsEnvironment, FileFormatDataSourceStats stats, ParquetReaderConfig config, HiveConfig hiveConfig)

Method Summary

Modifier and Type	Method	Description
Optional<ReaderPageSource>	createPageSource(org.apache.hadoop.conf.Configuration configuration, ConnectorSession session, org.apache.hadoop.fs.Path path, long start, long length, long estimatedFileSize, Properties schema, List<HiveColumnHandle> columns, TupleDomain<HiveColumnHandle> effectivePredicate, Optional<AcidInfo> acidInfo, OptionalInt bucketNumber, boolean originalFile, AcidTransaction transaction)
static ReaderPageSource	createPageSource(org.apache.hadoop.fs.Path path, long start, long length, long estimatedFileSize, List<HiveColumnHandle> columns, TupleDomain<HiveColumnHandle> effectivePredicate, boolean useColumnNames, HdfsEnvironment hdfsEnvironment, org.apache.hadoop.conf.Configuration configuration, ConnectorIdentity identity, org.joda.time.DateTimeZone timeZone, FileFormatDataSourceStats stats, ParquetReaderOptions options)	This method is available for other callers to use directly.
static Optional<org.apache.parquet.schema.Type>	getColumnType(HiveColumnHandle column, org.apache.parquet.schema.MessageType messageType, boolean useParquetColumnNames)
static TupleDomain<org.apache.parquet.column.ColumnDescriptor>	getParquetTupleDomain(Map<List<String>,RichColumnDescriptor> descriptorsByPath, TupleDomain<HiveColumnHandle> effectivePredicate, org.apache.parquet.schema.MessageType fileSchema, boolean useColumnNames)
static Optional<org.apache.parquet.schema.Type>	getParquetType(org.apache.parquet.schema.GroupType groupType, boolean useParquetColumnNames, HiveColumnHandle column)

Methods inherited from class java.lang.Object

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

Field Detail

PARQUET_ROW_INDEX_COLUMN

public static final HiveColumnHandle PARQUET_ROW_INDEX_COLUMN

If this object is passed as one of the columns for createPageSource, it will be populated as an additional column containing the index of each row read.

Constructor Detail

ParquetPageSourceFactory

@Inject
public ParquetPageSourceFactory(HdfsEnvironment hdfsEnvironment,
                                FileFormatDataSourceStats stats,
                                ParquetReaderConfig config,
                                HiveConfig hiveConfig)

Method Detail

createPageSource

public Optional<ReaderPageSource> createPageSource(org.apache.hadoop.conf.Configuration configuration,
                                                   ConnectorSession session,
                                                   org.apache.hadoop.fs.Path path,
                                                   long start,
                                                   long length,
                                                   long estimatedFileSize,
                                                   Properties schema,
                                                   List<HiveColumnHandle> columns,
                                                   TupleDomain<HiveColumnHandle> effectivePredicate,
                                                   Optional<AcidInfo> acidInfo,
                                                   OptionalInt bucketNumber,
                                                   boolean originalFile,
                                                   AcidTransaction transaction)

Specified by:

createPageSource in interface HivePageSourceFactory

createPageSource

public static ReaderPageSource createPageSource(org.apache.hadoop.fs.Path path,
                                                long start,
                                                long length,
                                                long estimatedFileSize,
                                                List<HiveColumnHandle> columns,
                                                TupleDomain<HiveColumnHandle> effectivePredicate,
                                                boolean useColumnNames,
                                                HdfsEnvironment hdfsEnvironment,
                                                org.apache.hadoop.conf.Configuration configuration,
                                                ConnectorIdentity identity,
                                                org.joda.time.DateTimeZone timeZone,
                                                FileFormatDataSourceStats stats,
                                                ParquetReaderOptions options)

This method is available for other callers to use directly.

getParquetType

public static Optional<org.apache.parquet.schema.Type> getParquetType(org.apache.parquet.schema.GroupType groupType,
                                                                      boolean useParquetColumnNames,
                                                                      HiveColumnHandle column)

getColumnType

public static Optional<org.apache.parquet.schema.Type> getColumnType(HiveColumnHandle column,
                                                                     org.apache.parquet.schema.MessageType messageType,
                                                                     boolean useParquetColumnNames)

getParquetTupleDomain

public static TupleDomain<org.apache.parquet.column.ColumnDescriptor> getParquetTupleDomain(Map<List<String>,RichColumnDescriptor> descriptorsByPath,
                                                                                            TupleDomain<HiveColumnHandle> effectivePredicate,
                                                                                            org.apache.parquet.schema.MessageType fileSchema,
                                                                                            boolean useColumnNames)