org.apache.drill.exec.physical.impl.join

Class LateralJoinBatch

java.lang.Object

org.apache.drill.exec.record.AbstractRecordBatch<T>

org.apache.drill.exec.record.AbstractBinaryRecordBatch<LateralJoinPOP>

org.apache.drill.exec.physical.impl.join.LateralJoinBatch

All Implemented Interfaces:

AutoCloseable, Iterable<VectorWrapper<?>>, LateralContract, CloseableRecordBatch, RecordBatch, VectorAccessible

public class LateralJoinBatch
extends AbstractBinaryRecordBatch<LateralJoinPOP>
implements LateralContract

RecordBatch implementation for the lateral join operator. Currently it's expected LATERAL to co-exists with UNNEST operator. Both LATERAL and UNNEST will share a contract with each other defined at LateralContract.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.drill.exec.record.AbstractBinaryRecordBatch

AbstractBinaryRecordBatch.Metric

Nested classes/interfaces inherited from class org.apache.drill.exec.record.AbstractRecordBatch

AbstractRecordBatch.BatchState

Nested classes/interfaces inherited from interface org.apache.drill.exec.record.RecordBatch

RecordBatch.IterOutcome

Field Summary

Fields inherited from class org.apache.drill.exec.record.AbstractBinaryRecordBatch

batchMemoryManager, left, LEFT_INDEX, leftUpstream, numInputs, right, RIGHT_INDEX, rightUpstream

Fields inherited from class org.apache.drill.exec.record.AbstractRecordBatch

batchStatsContext, container, context, oContext, popConfig, state, stats, unionTypeEnabled

Fields inherited from interface org.apache.drill.exec.record.RecordBatch

MAX_BATCH_ROW_COUNT

Constructor Summary

Constructors

Constructor and Description
LateralJoinBatch(LateralJoinPOP popConfig, FragmentContext context, RecordBatch left, RecordBatch right)

Method Summary

Modifier and Type	Method and Description
protected void	buildSchema() Prefetch a batch from left and right branch to know about the schema of each side.
protected void	cancelIncoming()
void	close()
void	dump() Perform dump of this batch's state to logs.
RecordBatch	getIncoming() Returns the left side incoming for the Lateral Join.
RecordBatch.IterOutcome	getLeftOutcome() Returns the current RecordBatch.IterOutcome for the left incoming batch
int	getRecordCount() Get the number of records.
int	getRecordIndex() Returns the current row index which the calling operator should process in current incoming left record batch.
RecordBatch.IterOutcome	innerNext() Gets the left and right incoming batch and produce the output batch.
protected boolean	prefetchFirstBatchFromBothSides() Get the left and right batch during build schema phase for LateralJoinBatch.
void	setMaxOutputRowCount(int outputRowCount) Used only for testing for cases when multiple output batches are produced for same input set
void	setUseMemoryManager(boolean useMemoryManager) Used only for testing to disable output batch calculation using memory manager and instead use the static max value set by setMaxOutputRowCount(int)

Methods inherited from class org.apache.drill.exec.record.AbstractBinaryRecordBatch

checkForEarlyFinish, getBatchMemoryManager, updateBatchMemoryManagerStats, verifyOutcomeToSetBatchState

Methods inherited from class org.apache.drill.exec.record.AbstractRecordBatch

cancel, checkContinue, getContainer, getContext, getOutgoingContainer, getPopConfig, getRecordBatchStatsContext, getSchema, getSelectionVector2, getSelectionVector4, getValueAccessorById, getValueVectorId, getWritableBatch, isRecordBatchStatsLoggingEnabled, iterator, next, next, next, schemaChangeException, schemaChangeException

Methods inherited from class java.lang.Object

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

Methods inherited from interface java.lang.Iterable

forEach, spliterator

Constructor Detail

LateralJoinBatch

public LateralJoinBatch(LateralJoinPOP popConfig,
                        FragmentContext context,
                        RecordBatch left,
                        RecordBatch right)
                 throws OutOfMemoryException

Throws:

OutOfMemoryException

Method Detail

innerNext

public RecordBatch.IterOutcome innerNext()

Gets the left and right incoming batch and produce the output batch. If the left incoming batch is empty then next on right branch is not called and empty batch with correct outcome is returned. If non empty left incoming batch is received then it call's next on right branch to get an incoming and finally produces output.

Specified by:

innerNext in class AbstractRecordBatch<LateralJoinPOP>

Returns:

IterOutcome state of the lateral join batch

close

public void close()

Specified by:

close in interface AutoCloseable

Overrides:

close in class AbstractRecordBatch<LateralJoinPOP>

getRecordCount

public int getRecordCount()

Description copied from interface: VectorAccessible

Get the number of records.

Specified by:

getRecordCount in interface VectorAccessible

Returns:

number of records

getIncoming

public RecordBatch getIncoming()

Returns the left side incoming for the Lateral Join. Used by right branch leaf operator of Lateral to process the records at leftJoinIndex.

Specified by:

getIncoming in interface LateralContract

Returns:

- RecordBatch received as left side incoming

getRecordIndex

public int getRecordIndex()

Returns the current row index which the calling operator should process in current incoming left record batch. LATERAL should never return it as -1 since that indicated current left batch is empty and LATERAL will never call next on right side with empty left batch

Specified by:

getRecordIndex in interface LateralContract

Returns:

- int - index of row to process.

getLeftOutcome

public RecordBatch.IterOutcome getLeftOutcome()

Returns the current RecordBatch.IterOutcome for the left incoming batch

Specified by:

getLeftOutcome in interface LateralContract

Returns:

The current outcome of left incoming batch.

prefetchFirstBatchFromBothSides

protected boolean prefetchFirstBatchFromBothSides()

Get the left and right batch during build schema phase for LateralJoinBatch. If left batch sees a failure outcome then we don't even call next on right branch, since there is no left incoming.

Overrides:

prefetchFirstBatchFromBothSides in class AbstractBinaryRecordBatch<LateralJoinPOP>

Returns:

true if both the left/right batch was received without failure outcome. false if either of batch is received with failure outcome.

buildSchema

protected void buildSchema()

Prefetch a batch from left and right branch to know about the schema of each side. Then adds value vector in output container based on those schemas. For this phase LATERAL always expect's an empty batch from right side which UNNEST should abide by.

Overrides:

buildSchema in class AbstractRecordBatch<LateralJoinPOP>

Throws:

SchemaChangeException - if batch schema was changed during execution

cancelIncoming

protected void cancelIncoming()

Overrides:

cancelIncoming in class AbstractBinaryRecordBatch<LateralJoinPOP>

setMaxOutputRowCount

public void setMaxOutputRowCount(int outputRowCount)

Used only for testing for cases when multiple output batches are produced for same input set

Parameters:

outputRowCount - - Max rows that output batch can hold

setUseMemoryManager

public void setUseMemoryManager(boolean useMemoryManager)

Used only for testing to disable output batch calculation using memory manager and instead use the static max value set by setMaxOutputRowCount(int)

Parameters:

useMemoryManager - - false - disable memory manager update to take effect, true enable memory manager update

dump

public void dump()

Description copied from interface: RecordBatch

Perform dump of this batch's state to logs.

Specified by:

dump in interface RecordBatch