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

Class ContainerState

java.lang.Object

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

Direct Known Subclasses:

ListState, RepeatedListState, TupleState, UnionState

public abstract class ContainerState
extends Object

Abstract representation of a container of vectors: a row, a map, a repeated map, a list or a union.

The container is responsible for creating new columns in response from a writer listener event. Column creation requires a set of four items:

• The value vector (which may be null if the column is not projected.
• The writer for the column.
• A vector state that manages allocation, overflow, cleanup and other vector-specific tasks.
• A column state which orchestrates the above three items.

Field Summary

Fields

Modifier and Type	Field and Description
protected org.apache.drill.exec.physical.resultSet.impl.LoaderInternals	loader
protected ColumnState	parentColumn
protected ProjectionFilter	projectionSet
protected ResultVectorCache	vectorCache Vector cache for this loader.

Constructor Summary

Constructors

Constructor and Description
ContainerState(org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader, ResultVectorCache vectorCache)
ContainerState(org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader, ResultVectorCache vectorCache, ProjectionFilter projectionSet)

Method Summary

Modifier and Type	Method and Description
ColumnState	addColumn(ColumnMetadata columnSchema)
protected abstract void	addColumn(ColumnState colState)
void	bindColumnState(ColumnState parentState)
void	close() Clean up state (such as backup vectors) associated with the state for each vector.
protected abstract Collection<ColumnState>	columnStates()
void	harvestWithLookAhead() Writing of a row batch is complete, and an overflow occurred.
abstract int	innerCardinality()
protected abstract boolean	isVersioned() Reports whether this container is subject to version management.
protected org.apache.drill.exec.physical.resultSet.impl.LoaderInternals	loader()
protected ProjectionFilter	projection()
void	rollover() A column within the row batch overflowed.
void	startBatch(boolean schemaOnly) Start a new batch by shifting the overflow buffers back into the main write vectors and updating the writers.
void	updateCardinality() In order to allocate the correct-sized vectors, the container must know its member cardinality: the number of elements in each row.
ResultVectorCache	vectorCache()

Methods inherited from class java.lang.Object

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

Field Detail

loader

protected final org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader

projectionSet

protected final ProjectionFilter projectionSet

parentColumn

protected ColumnState parentColumn

vectorCache

protected final ResultVectorCache vectorCache

Vector cache for this loader. ResultSetOptionBuilder#setVectorCache().

Constructor Detail

ContainerState

public ContainerState(org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader,
                      ResultVectorCache vectorCache,
                      ProjectionFilter projectionSet)

ContainerState

public ContainerState(org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader,
                      ResultVectorCache vectorCache)

Method Detail

bindColumnState

public void bindColumnState(ColumnState parentState)

innerCardinality

public abstract int innerCardinality()

addColumn

protected abstract void addColumn(ColumnState colState)

columnStates

protected abstract Collection<ColumnState> columnStates()

projection

protected ProjectionFilter projection()

isVersioned

protected abstract boolean isVersioned()

Reports whether this container is subject to version management. Version management adds columns to the output container at harvest time based on whether they should appear in the output batch.

Returns:

true if versioned

loader

protected org.apache.drill.exec.physical.resultSet.impl.LoaderInternals loader()

vectorCache

public ResultVectorCache vectorCache()

addColumn

public ColumnState addColumn(ColumnMetadata columnSchema)

updateCardinality

public void updateCardinality()

In order to allocate the correct-sized vectors, the container must know its member cardinality: the number of elements in each row. This is 1 for a single map or union, but may be any number for a map array or a list. Then, this value is recursively pushed downward to compute the cardinality of lists of maps that contains lists of maps, and so on.

startBatch

public void startBatch(boolean schemaOnly)

Start a new batch by shifting the overflow buffers back into the main write vectors and updating the writers.

rollover

public void rollover()

A column within the row batch overflowed. Prepare to absorb the rest of the in-flight row by rolling values over to a new vector, saving the complete vector for later. This column could have a value for the overflow row, or for some previous row, depending on exactly when and where the overflow occurs.

harvestWithLookAhead

public void harvestWithLookAhead()

Writing of a row batch is complete, and an overflow occurred. Prepare the vector for harvesting to send downstream. Set aside the look-ahead vector and put the full vector buffer back into the active vector.

close

public void close()

Clean up state (such as backup vectors) associated with the state for each vector.