Package org.apache.drill.exec.physical.resultSet.impl

Interface VectorState

All Known Implementing Classes:
ListState.ListVectorState, NullableVectorState, NullVectorState, NullVectorState.UnmanagedVectorState, RepeatedListState.RepeatedListVectorState, RepeatedVectorState, SingleVectorState, SingleVectorState.FixedWidthVectorState, SingleVectorState.IsSetVectorState, SingleVectorState.OffsetVectorState, SingleVectorState.SimpleVectorState, SingleVectorState.VariableWidthVectorState, TupleState.DictArrayVectorState, TupleState.DictVectorState, TupleState.MapVectorState, TupleState.SingleDictVectorState, UnionState.UnionVectorState
public interface VectorState
Handles batch and overflow operation for a (possibly compound) vector.

The data model is the following:

  • Column model
    • Value vector itself
    • Column writer
    • Column schema
    • Column coordinator (this class)
The vector state coordinates events between the result set loader on the one side and the vectors, writers and schema on the other. For example: 

 Result Set       Vector
   Loader   invalid input: '<'-->  State   invalid input: '<'-->    Vectors
Events from the row set loader deal with allocation, roll-over, harvesting completed batches and so on. Events from the writer, via the tuple model deal with adding columns and column overflow.

  • Method Summary

    Modifier and Type
    Method
    Description
    int
    allocate(int cardinality)
    Allocate a new vector with the number of elements given.
    void
    close()
    Clear the vector(s) associated with this state.
    void
    dump(HierarchicalFormatter format)
     
    void
    harvestWithLookAhead()
    A batch is being harvested after an overflow.
    boolean
    isProjected()
    Report whether this column is projected (has materialized vectors), or is unprojected (has no materialized backing.)
    void
    rollover(int cardinality)
    A vector has overflowed.
    void
    startBatchWithLookAhead()
    A new batch is starting while an look-ahead vector exists.
    <T extends ValueVector>
    T
    vector()
    Underlying vector: the one presented to the consumer of the result set loader.

  • Method Details

    • allocate

      int allocate(int cardinality)
      Allocate a new vector with the number of elements given. If the vector is an array, then the cardinality given is the number of arrays.
      Parameters:
      cardinality - number of elements desired in the allocated vector
      Returns:
      the number of bytes allocated

    • rollover

      void rollover(int cardinality)
      A vector has overflowed. Create a new look-ahead vector of the given cardinality, then copy the overflow values from the main vector to the look-ahead vector.
      Parameters:
      cardinality - the number of elements in the new vector. If this vector is an array, then this is the number of arrays

    • harvestWithLookAhead

      void harvestWithLookAhead()
      A batch is being harvested after an overflow. Put the full batch back into the main vector so it can be harvested.

    • startBatchWithLookAhead

      void startBatchWithLookAhead()
      A new batch is starting while an look-ahead vector exists. Move the look-ahead buffers into the main vector to prepare for writing the rest of the batch.

    • close

      void close()
      Clear the vector(s) associated with this state.

    • vector

      <T extends ValueVector> T vector()
      Underlying vector: the one presented to the consumer of the result set loader.

    • isProjected

      boolean isProjected()
      Report whether this column is projected (has materialized vectors), or is unprojected (has no materialized backing.)
      Returns:
      true if the column is projected to the output, false if not

    • dump

      void dump(HierarchicalFormatter format)