org.infinispan

Interface LockedStream<K,V>

All Superinterfaces:

AutoCloseable, BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>, BaseStream<CacheEntry<K,V>,LockedStream<K,V>>

All Known Implementing Classes:

LockedStreamImpl, TxLockedStreamImpl

public interface LockedStream<K,V>
extends BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Stream that allows for operation upon data solely with side effects by using forEach(BiConsumer) where the BiConsumer is invoked while guaranteeing that the entry being passed is properly locked for the entire duration of the invocation.

An attempt is made to acquire the lock for an entry using the default LockingConfiguration.lockAcquisitionTimeout() before invoking any operations on it.

Since:

9.1

Author:

wburns

Nested Class Summary

Nested classes/interfaces inherited from interface org.infinispan.BaseCacheStream

BaseCacheStream.SegmentCompletionListener

Method Summary

Modifier and Type	Method and Description
LockedStream<K,V>	disableRehashAware() Disables tracking of rehash events that could occur to the underlying cache.
LockedStream<K,V>	distributedBatchSize(int batchSize) Controls how many keys are returned from a remote node when using a stream terminal operation with a distributed cache to back this stream.
LockedStream<K,V>	filter(Predicate<? super CacheEntry<K,V>> predicate) Returns a locked stream consisting of the elements of this stream that match the given predicate.
default LockedStream<K,V>	filter(SerializablePredicate<? super CacheEntry<K,V>> predicate) Same as filter(Predicate) except that the Predicate must also implement Serializable
LockedStream<K,V>	filterKeys(Set<?> keys) Filters which entries are returned by only returning ones that map to the given key.
LockedStream<K,V>	filterKeySegments(IntSet segments) Filters which entries are returned by what segment they are present in.
LockedStream<K,V>	filterKeySegments(Set<Integer> segments) Deprecated. This is to be replaced by filterKeySegments(IntSet)
void	forEach(BiConsumer<Cache<K,V>,? super CacheEntry<K,V>> biConsumer) Performs an action for each element of this stream on the primary owner of the given key.
default void	forEach(SerializableBiConsumer<Cache<K,V>,? super CacheEntry<K,V>> biConsumer) Same as forEach(BiConsumer) except that the BiConsumer must also implement Serializable
<R> Map<K,R>	invokeAll(BiFunction<Cache<K,V>,? super CacheEntry<K,V>,R> biFunction) Performs a BiFunction for each element of this stream on the primary owner of each entry returning a value.
default <R> Map<K,R>	invokeAll(SerializableBiFunction<Cache<K,V>,? super CacheEntry<K,V>,R> biFunction) Same as invokeAll(BiFunction) except that the BiFunction must also implement Serializable
Iterator<CacheEntry<K,V>>	iterator() This method is not supported when using a LockedStream
LockedStream<K,V>	parallelDistribution() This would enable sending requests to all other remote nodes when a terminal operator is performed.
LockedStream<K,V>	segmentCompletionListener(BaseCacheStream.SegmentCompletionListener listener) This method is not supported when using a LockedStream
LockedStream<K,V>	sequentialDistribution() This would disable sending requests to all other remote nodes compared to one at a time.
Spliterator<CacheEntry<K,V>>	spliterator() This method is not supported when using a LockedStream
LockedStream<K,V>	timeout(long time, TimeUnit unit) Sets the timeout for the acquisition of the lock for each entry.

Methods inherited from interface java.util.stream.BaseStream

close, isParallel, onClose, parallel, sequential, unordered

Method Detail

filter

LockedStream<K,V> filter(Predicate<? super CacheEntry<K,V>> predicate)

Returns a locked stream consisting of the elements of this stream that match the given predicate.

This filter is after the lock is acquired for the given key. This way the filter will see the same value as the consumer is given.

Parameters:

predicate - predicate

Returns:

a LockedStream with the filter applied

filter

default LockedStream<K,V> filter(SerializablePredicate<? super CacheEntry<K,V>> predicate)

Same as filter(Predicate) except that the Predicate must also implement Serializable

The compiler will pick this overload for lambda parameters, making them Serializable

Parameters:

predicate - the predicate to filter out unwanted entries

Returns:

a LockedStream with the filter applied

forEach

void forEach(BiConsumer<Cache<K,V>,? super CacheEntry<K,V>> biConsumer)

Performs an action for each element of this stream on the primary owner of the given key.

This method is performed while holding exclusive lock over the given entry and will be released only after the consumer has completed. In the function, entry.setValue(newValue) is equivalent to cache.put(entry.getKey(), newValue).

If using pessimistic transactions this lock is not held using a transaction and thus the user can start a transaction in this consumer which also must be completed before returning. A transaction can be started in the consumer and if done it will share the same lock used to obtain the key.

Remember that if you are using an explicit transaction or an async method that these must be completed before the consumer returns to guarantee that they are operating within the scope of the lock for the given key. Failure to do so will lead into possible inconsistency as they will be performing operations without the proper locking.

Some methods on the provided cache may not work as expected. These include Cache.putForExternalRead(Object, Object), AdvancedCache.lock(Object[]), AdvancedCache.lock(Collection), and AdvancedCache.removeGroup(String). If these methods are used inside of the Consumer on the cache it will throw a IllegalStateException. This is due to possible interactions with locks while using these commands.

Parameters:

biConsumer - the biConsumer to run for each entry under their lock

forEach

default void forEach(SerializableBiConsumer<Cache<K,V>,? super CacheEntry<K,V>> biConsumer)

Same as forEach(BiConsumer) except that the BiConsumer must also implement Serializable

The compiler will pick this overload for lambda parameters, making them Serializable

Parameters:

biConsumer - the biConsumer to run for each entry under their lock

invokeAll

@Experimental
<R> Map<K,R> invokeAll(BiFunction<Cache<K,V>,? super CacheEntry<K,V>,R> biFunction)

Performs a BiFunction for each element of this stream on the primary owner of each entry returning a value. The returned value from the function will be sent back to the user mapped to the key that generated it, with all of these stored in a map. Both the BiFunction and the returned value must be Serializable in some way. This method will return only after all entries have been processed.

This method is currently marked as Experimental since this method returns a Map and requires blocking. This operation could take a deal of time and as such should be done using an asynchronous API. Most likely this return type will be changed to use some sort of asynchronous return value. This method is here until this can be implemented.

This BiFunction is invoked while holding an exclusive lock over the given entry that will be released only after the function has completed. In the function, entry.setValue(newValue) is equivalent to cache.put(entry.getKey(), newValue).

If using pessimistic transactions this lock is not held using a transaction and thus the user can start a transaction in this consumer which also must be completed before returning. A transaction can be started in the biFunction and if done it will share the same lock used to obtain the key.

Remember if you are using an explicit transaction or an async method that these must be completed before the consumer returns to guarantee that they are operating within the scope of the lock for the given key. Failure to do so will lead into possible inconsistency as they will be performing operations without the proper locking.

Some methods on the provided cache may not work as expected. These include Cache.putForExternalRead(Object, Object), AdvancedCache.lock(Object[]), AdvancedCache.lock(Collection), and AdvancedCache.removeGroup(String). If these methods are used inside of the Consumer on the cache it will throw a IllegalStateException. This is due to possible interactions with locks while using these commands.

Type Parameters:

R - the return type

Parameters:

biFunction - the biFunction to run for each entry under their lock

Returns:

a map with each key mapped to the value returned from the bi function

invokeAll

@Experimental
default <R> Map<K,R> invokeAll(SerializableBiFunction<Cache<K,V>,? super CacheEntry<K,V>,R> biFunction)

Same as invokeAll(BiFunction) except that the BiFunction must also implement Serializable

The compiler will pick this overload for lambda parameters, making them Serializable

Type Parameters:

R - the return type

Parameters:

biFunction - the biFunction to run for each entry under their lock

Returns:

a map with each key mapped to the value returned from the bi function

sequentialDistribution

LockedStream<K,V> sequentialDistribution()

This would disable sending requests to all other remote nodes compared to one at a time. This can reduce memory pressure on the originator node at the cost of performance.

Parallel distribution is enabled by default except for CacheStream.iterator() and CacheStream.spliterator()

Specified by:

sequentialDistribution in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Returns:

a stream with parallel distribution disabled

parallelDistribution

LockedStream<K,V> parallelDistribution()

This would enable sending requests to all other remote nodes when a terminal operator is performed. This requires additional overhead as it must process results concurrently from various nodes, but should perform faster in the majority of cases.

Parallel distribution is enabled by default except for CacheStream.iterator() and CacheStream.spliterator()

Specified by:

parallelDistribution in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Returns:

a stream with parallel distribution enabled.

filterKeySegments

LockedStream<K,V> filterKeySegments(Set<Integer> segments)

Deprecated. This is to be replaced by filterKeySegments(IntSet)

Filters which entries are returned by what segment they are present in. This method can be substantially more efficient than using a regular CacheStream.filter(Predicate) method as this can control what nodes are asked for data and what entries are read from the underlying CacheStore if present.

Specified by:

filterKeySegments in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

segments - The segments to use for this stream operation. Any segments not in this set will be ignored.

Returns:

a stream with the segments filtered.

filterKeySegments

LockedStream<K,V> filterKeySegments(IntSet segments)

Filters which entries are returned by what segment they are present in. This method can be substantially more efficient than using a regular CacheStream.filter(Predicate) method as this can control what nodes are asked for data and what entries are read from the underlying CacheStore if present.

Specified by:

filterKeySegments in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

segments - The segments to use for this stream operation. Any segments not in this set will be ignored.

Returns:

a stream with the segments filtered.

filterKeys

LockedStream<K,V> filterKeys(Set<?> keys)

Filters which entries are returned by only returning ones that map to the given key. This method will be faster than a regular CacheStream.filter(Predicate) if the filter is holding references to the same keys.

Specified by:

filterKeys in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

keys - The keys that this stream will only operate on.

Returns:

a stream with the keys filtered.

distributedBatchSize

LockedStream<K,V> distributedBatchSize(int batchSize)

Controls how many keys are returned from a remote node when using a stream terminal operation with a distributed cache to back this stream. This value is ignored when terminal operators that don't track keys are used. Key tracking terminal operators are CacheStream.iterator(), CacheStream.spliterator(), CacheStream.forEach(Consumer). Please see those methods for additional information on how this value may affect them.

This value may be used in the case of a a terminal operator that doesn't track keys if an intermediate operation is performed that requires bringing keys locally to do computations. Examples of such intermediate operations are CacheStream.sorted(), CacheStream.sorted(Comparator), CacheStream.distinct(), CacheStream.limit(long), CacheStream.skip(long)

This value is always ignored when this stream is backed by a cache that is not distributed as all values are already local.

Specified by:

distributedBatchSize in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

batchSize - The size of each batch. This defaults to the state transfer chunk size.

Returns:

a stream with the batch size updated

disableRehashAware

LockedStream<K,V> disableRehashAware()

Disables tracking of rehash events that could occur to the underlying cache. If a rehash event occurs while a terminal operation is being performed it is possible for some values that are in the cache to not be found. Note that you will never have an entry duplicated when rehash awareness is disabled, only lost values.

Most terminal operations will run faster with rehash awareness disabled even without a rehash occuring. However if a rehash occurs with this disabled be prepared to possibly receive only a subset of values.

Specified by:

disableRehashAware in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Returns:

a stream with rehash awareness disabled.

timeout

LockedStream<K,V> timeout(long time,
                          TimeUnit unit)

Sets the timeout for the acquisition of the lock for each entry.

Specified by:

timeout in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

time - the maximum time to wait

unit - the time unit of the timeout argument

Returns:

a LockedStream with the timeout applied

segmentCompletionListener

LockedStream<K,V> segmentCompletionListener(BaseCacheStream.SegmentCompletionListener listener)
                                     throws UnsupportedOperationException

This method is not supported when using a LockedStream

Specified by:

segmentCompletionListener in interface BaseCacheStream<CacheEntry<K,V>,LockedStream<K,V>>

Parameters:

listener - The listener that will be called back as segments are completed.

Returns:

a stream with the listener registered.

Throws:

UnsupportedOperationException

iterator

Iterator<CacheEntry<K,V>> iterator()
                            throws UnsupportedOperationException

This method is not supported when using a LockedStream

Specified by:

iterator in interface BaseStream<CacheEntry<K,V>,LockedStream<K,V>>

Throws:

UnsupportedOperationException

spliterator

Spliterator<CacheEntry<K,V>> spliterator()
                                  throws UnsupportedOperationException

This method is not supported when using a LockedStream

Specified by:

spliterator in interface BaseStream<CacheEntry<K,V>,LockedStream<K,V>>

Throws:

UnsupportedOperationException