Package org.apache.drill.exec.vector.accessor

Interface ObjectWriter

All Superinterfaces:
ColumnWriter
All Known Implementing Classes:
AbstractArrayWriter.ArrayObjectWriter, AbstractObjectWriter, AbstractScalarWriterImpl.ScalarObjectWriter, AbstractTupleWriter.TupleObjectWriter, DictEntryWriter.DictEntryObjectWriter, ObjectDictWriter.DictObjectWriter, UnionWriterImpl.VariantObjectWriter
public interface ObjectWriter extends ColumnWriter
Represents a column within a tuple. A column can be an array, a scalar or a tuple. Each has an associated column metadata (schema) and a writer. The writer is one of three kinds, depending on the kind of the column. If the column is a map, then the column also has an associated tuple loader to define and write to the tuple.

This interface defines a writer to set values for value vectors using a simple, uniform interface modeled after a JSON object. Every column value is an object of one of three types: scalar, array or tuple. Methods exist to "cast" this object to the proper type. This model allows a very simple representation: tuples (rows, maps) consist of objects. Arrays are lists of objects.

Every column resides at an index, is defined by a schema, is backed by a value vector, and and is written to by a writer. Each column also tracks the schema version in which it was added to detect schema evolution. Each column has an optional overflow vector that holds overflow record values when a batch becomes full.

  • Method Details

    • scalar

      ScalarWriter scalar()

    • tuple

      TupleWriter tuple()

    • array

      ArrayWriter array()

    • variant

      VariantWriter variant()

    • dict

      DictWriter dict()

    • writer

      ColumnWriter writer()
      Generic version of the above, for dynamic handling of writers.
      Returns:
      the generic form of the column writer

    • events

      WriterEvents events()
      The internal state behind this writer. To be used only by the implementation, not by the client.