Package org.apache.drill.exec.store.parquet.columnreaders

Class Summary

Class	Description
BatchReader	Base strategy for reading a batch of Parquet records.
BatchReader.FixedWidthReader	Strategy for reading a record batch when all columns are fixed-width.
BatchReader.MockBatchReader	Strategy for reading mock records.
BatchReader.VariableWidthReader	Strategy for reading a record batch when at last one column is variable width.
ColumnReader<V extends ValueVector>
ColumnReaderFactory
FixedWidthRepeatedReader
NullableFixedByteAlignedReaders
NullableFixedByteAlignedReaders.CorruptionDetectingNullableDateReader	Old versions of Drill were writing a non-standard format for date.
NullableFixedByteAlignedReaders.NullableCorruptDateReader	Old versions of Drill were writing a non-standard format for date.
NullableFixedByteAlignedReaders.NullableDateReader
NullableFixedByteAlignedReaders.NullableIntervalReader
NullableVarLengthValuesColumn<V extends ValueVector>
ParquetColumnMetadata	Represents a single column read from the Parquet file by the record reader.
ParquetFixedWidthDictionaryReaders
ParquetRecordReader
ParquetSchema	Mapping from the schema of the Parquet file to that of the record reader to the schema that Drill and the Parquet reader uses.
ParquetToDrillTypeConverter
ReadState	Internal state for reading from a Parquet file.
VarLenBinaryReader	Class which handles reading a batch of rows from a set of variable columns
VarLenColumnBulkInput<V extends ValueVector>	Implements the VarLenBulkInput interface to optimize data copy
VarLengthColumn<V extends ValueVector>
VarLengthColumnReaders
VarLengthColumnReaders.NullableVarBinaryColumn
VarLengthColumnReaders.NullableVarCharColumn
VarLengthColumnReaders.NullableVarDecimalColumn
VarLengthColumnReaders.VarBinaryColumn
VarLengthColumnReaders.VarCharColumn
VarLengthColumnReaders.VarDecimalColumn
VarLengthValuesColumn<V extends ValueVector>
VarLenOverflowReader	This class is responsible for processing serialized overflow data (generated in a previous batch); this way overflow data becomes an input source and is thus a) efficiently re-loaded into the current batch ValueVector and b) subjected to the same batching constraints rules.