com.hazelcast.map.impl.query

Class ParallelPartitionScanExecutor

java.lang.Object

com.hazelcast.map.impl.query.ParallelPartitionScanExecutor

All Implemented Interfaces:

PartitionScanExecutor

public class ParallelPartitionScanExecutor
extends Object
implements PartitionScanExecutor

Implementation of the PartitionScanExecutor which executes the partition scan in a parallel-fashion delegating to the underlying executor.

Constructor Summary

Constructors

Constructor and Description
ParallelPartitionScanExecutor(PartitionScanRunner partitionScanRunner, ManagedExecutorService executor, int timeoutInMillis)

Method Summary

Modifier and Type	Method and Description
List<QueryableEntry>	execute(String mapName, Predicate predicate, Collection<Integer> partitions)
QueryableEntriesSegment	execute(String mapName, Predicate predicate, int partitionId, int tableIndex, int fetchSize) Executes the predicate on a partition chunk.
protected Future<Collection<QueryableEntry>>	runPartitionScanForPartition(String name, Predicate predicate, int partitionId)
protected List<QueryableEntry>	runUsingPartitionScanWithoutPaging(String name, Predicate predicate, Collection<Integer> partitions)

Methods inherited from class java.lang.Object

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

Constructor Detail

ParallelPartitionScanExecutor

public ParallelPartitionScanExecutor(PartitionScanRunner partitionScanRunner,
                                     ManagedExecutorService executor,
                                     int timeoutInMillis)

Method Detail

execute

public List<QueryableEntry> execute(String mapName,
                                    Predicate predicate,
                                    Collection<Integer> partitions)

Specified by:

execute in interface PartitionScanExecutor

execute

public QueryableEntriesSegment execute(String mapName,
                                       Predicate predicate,
                                       int partitionId,
                                       int tableIndex,
                                       int fetchSize)

Executes the predicate on a partition chunk. The offset in the partition is defined by the tableIndex and the soft limit is defined by the fetchSize. The method returns the matched entries and an index from which new entries can be fetched which allows for efficient iteration of query results.

NOTE Iterating the query results using the returned next table index should be done only when the IMap is not being mutated and the cluster is stable (there are no migrations or membership changes). In other cases, entries are rearranged and the you may get the same query result twice or may miss some query results that match the predicate. Parallel execution for a partition chunk query is not supported.

Specified by:

execute in interface PartitionScanExecutor

Parameters:

mapName - the map name

predicate - the predicate which the entries must match

partitionId - the partition which is queried

tableIndex - the index from which entries are queried

fetchSize - the soft limit for the number of entries to fetch

Returns:

entries matching the predicate and a table index from which new entries can be fetched

runUsingPartitionScanWithoutPaging

protected List<QueryableEntry> runUsingPartitionScanWithoutPaging(String name,
                                                                  Predicate predicate,
                                                                  Collection<Integer> partitions)

runPartitionScanForPartition

protected Future<Collection<QueryableEntry>> runPartitionScanForPartition(String name,
                                                                          Predicate predicate,
                                                                          int partitionId)