com.hazelcast.query.impl

Class Indexes

java.lang.Object

com.hazelcast.query.impl.Indexes

public class Indexes
extends Object

Contains all indexes for a data-structure, e.g. an IMap.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Indexes.Builder Builder which is used to create a new Indexes object.

Field Summary

Fields

Modifier and Type	Field and Description
static int	SKIP_PARTITIONS_COUNT_CHECK The partitions count check detects a race condition when a query is executed on the index which is under (re)construction.

Method Summary

Modifier and Type	Method and Description
InternalIndex	addOrGetIndex(String name, boolean ordered) Obtains the existing index or creates a new one (if an index doesn't exist yet) for the given name in this indexes instance.
void	clearAll() Clears contents of indexes managed by this instance.
void	createIndexesFromRecordedDefinitions() Creates indexes according to the index definitions stored inside this indexes.
void	destroyIndexes() Destroys and then removes all the indexes from this indexes instance.
InternalIndex[]	getCompositeIndexes() Returns all the composite indexes known to this indexes instance.
TypeConverter	getConverter(String attribute)
InternalIndex	getIndex(String name)
Collection<IndexInfo>	getIndexDefinitions()
InternalIndex[]	getIndexes() Returns all the indexes known to this indexes instance.
IndexesStats	getIndexesStats() Returns the indexes stats of this indexes instance.
boolean	haveAtLeastOneIndex() Returns true if this indexes instance contains at least one index, false otherwise.
boolean	haveAtLeastOneIndexOrDefinition() Returns true if the indexes instance contains either at least one index or its definition, false otherwise.
boolean	isGlobal()
static void	markPartitionAsIndexed(int partitionId, InternalIndex[] indexes) Marks the given partition as indexed by the given indexes.
static void	markPartitionAsUnindexed(int partitionId, InternalIndex[] indexes) Marks the given partition as unindexed by the given indexes.
InternalIndex	matchIndex(String pattern, Class<? extends Predicate> predicateClass, QueryContext.IndexMatchHint matchHint, int ownedPartitionCount) Matches an index for the given pattern and match hint that can evaluate the given predicate class.
InternalIndex	matchIndex(String pattern, QueryContext.IndexMatchHint matchHint, int ownedPartitionCount) Matches an index for the given pattern and match hint.
static Indexes.Builder	newBuilder(SerializationService ss, IndexCopyBehavior indexCopyBehavior)
void	putEntry(QueryableEntry queryableEntry, Object oldValue, Index.OperationSource operationSource) Inserts a new queryable entry into this indexes instance or updates the existing one.
Set<QueryableEntry>	query(Predicate predicate, int ownedPartitionCount) Performs a query on this indexes instance using the given predicate.
void	recordIndexDefinition(String name, boolean ordered) Records the given index definition in this indexes without creating an index.
void	removeEntry(Data key, Object value, Index.OperationSource operationSource) Removes the entry from this indexes instance identified by the given key and value.

Methods inherited from class java.lang.Object

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

Field Detail

SKIP_PARTITIONS_COUNT_CHECK

public static final int SKIP_PARTITIONS_COUNT_CHECK

The partitions count check detects a race condition when a query is executed on the index which is under (re)construction. The negative value means the check should be skipped.

See Also:

Constant Field Values

Method Detail

markPartitionAsIndexed

public static void markPartitionAsIndexed(int partitionId,
                                          InternalIndex[] indexes)

Marks the given partition as indexed by the given indexes.

Parameters:

partitionId - the ID of the partition to mark as indexed.

indexes - the indexes by which the given partition is indexed.

markPartitionAsUnindexed

public static void markPartitionAsUnindexed(int partitionId,
                                            InternalIndex[] indexes)

Marks the given partition as unindexed by the given indexes.

Parameters:

partitionId - the ID of the partition to mark as unindexed.

indexes - the indexes by which the given partition is unindexed.

newBuilder

public static Indexes.Builder newBuilder(SerializationService ss,
                                         IndexCopyBehavior indexCopyBehavior)

Parameters:

ss - the serializationService

indexCopyBehavior - the indexCopyBehavior

Returns:

new builder instance which will be used to create Indexes object.

See Also:

IndexCopyBehavior

addOrGetIndex

public InternalIndex addOrGetIndex(String name,
                                   boolean ordered)

Obtains the existing index or creates a new one (if an index doesn't exist yet) for the given name in this indexes instance.

Parameters:

name - the name of the index; the passed value might not represent a canonical index name (as specified by {@link Index#getName()), in this case the method canonicalizes it.

ordered - true if the new index should be ordered, false otherwise.

Returns:

the existing or created index.

recordIndexDefinition

public void recordIndexDefinition(String name,
                                  boolean ordered)

Records the given index definition in this indexes without creating an index.

Parameters:

name - the name of the index; the passed value might not represent a canonical index name (as specified by {@link Index#getName()), in this case the method canonicalizes it.

ordered - true if the new index should be ordered, false otherwise.

createIndexesFromRecordedDefinitions

public void createIndexesFromRecordedDefinitions()

Creates indexes according to the index definitions stored inside this indexes.

getIndexDefinitions

public Collection<IndexInfo> getIndexDefinitions()

getIndexes

public InternalIndex[] getIndexes()

Returns all the indexes known to this indexes instance.

getCompositeIndexes

public InternalIndex[] getCompositeIndexes()

Returns all the composite indexes known to this indexes instance.

destroyIndexes

public void destroyIndexes()

Destroys and then removes all the indexes from this indexes instance.

clearAll

public void clearAll()

Clears contents of indexes managed by this instance.

haveAtLeastOneIndex

public boolean haveAtLeastOneIndex()

Returns true if this indexes instance contains at least one index, false otherwise.

haveAtLeastOneIndexOrDefinition

public boolean haveAtLeastOneIndexOrDefinition()

Returns true if the indexes instance contains either at least one index or its definition, false otherwise.

Returns:

putEntry

public void putEntry(QueryableEntry queryableEntry,
                     Object oldValue,
                     Index.OperationSource operationSource)

Inserts a new queryable entry into this indexes instance or updates the existing one.

Parameters:

queryableEntry - the queryable entry to insert or update.

oldValue - the old entry value to update, null if inserting the new entry.

operationSource - the operation source.

removeEntry

public void removeEntry(Data key,
                        Object value,
                        Index.OperationSource operationSource)

Removes the entry from this indexes instance identified by the given key and value.

Parameters:

key - the key if the entry to remove.

value - the value of the entry to remove.

operationSource - the operation source.

isGlobal

public boolean isGlobal()

Returns:

true if the index is global-per map, meaning there is just a single instance of this object per map. Global indexes are used in on-heap maps, since they give a significant performance boost. The opposite of global indexes are partitioned-indexes which are stored locally per partition.

In case of a partitioned-index, each query has to query the index in each partition separately, which all-together may be around 3 times slower than querying a single global index.

getIndex

public InternalIndex getIndex(String name)

Returns:

the index with the given name or null if such index does not exist. It's a caller's responsibility to canonicalize the passed index name as specified by Index.getName().

query

public Set<QueryableEntry> query(Predicate predicate,
                                 int ownedPartitionCount)

Performs a query on this indexes instance using the given predicate.

Parameters:

predicate - the predicate to evaluate.

ownedPartitionCount - a count of owned partitions a query runs on. Negative value indicates that the value is not defined.

Returns:

the produced result set or null if the query can't be performed using the indexes known to this indexes instance.

matchIndex

public InternalIndex matchIndex(String pattern,
                                QueryContext.IndexMatchHint matchHint,
                                int ownedPartitionCount)

Matches an index for the given pattern and match hint.

Parameters:

pattern - the pattern to match an index for. May be either an attribute name or an exact index name.

matchHint - the match hint.

ownedPartitionCount - a count of owned partitions a query runs on. Negative value indicates that the value is not defined.

Returns:

the matched index or null if nothing matched.

See Also:

QueryContext.IndexMatchHint, QueryContext.matchIndex(java.lang.String, com.hazelcast.query.impl.QueryContext.IndexMatchHint)

matchIndex

public InternalIndex matchIndex(String pattern,
                                Class<? extends Predicate> predicateClass,
                                QueryContext.IndexMatchHint matchHint,
                                int ownedPartitionCount)

Matches an index for the given pattern and match hint that can evaluate the given predicate class.

Parameters:

pattern - the pattern to match an index for. May be either an attribute name or an exact index name.

predicateClass - the predicate class the matched index must be able to evaluate.

matchHint - the match hint.

ownedPartitionCount - a count of owned partitions a query runs on. Negative value indicates that the value is not defined.

Returns:

the matched index or null if nothing matched.

See Also:

QueryContext.IndexMatchHint, Index.evaluate(com.hazelcast.query.Predicate)

getConverter

public TypeConverter getConverter(String attribute)

Returns:

a converter instance for the given attribute or null if a converter is not available. The later may happen if the attribute is unknown to this indexes instance or there are no populated indexes involving the given attribute.

getIndexesStats

public IndexesStats getIndexesStats()

Returns the indexes stats of this indexes instance.