Interface CachePublisher<K,V>

Type Parameters:
K - key type in the cache
V - value type in the cache

This API is currently Experimental and may be changed or even removed later, use at your own risk.

All Known Implementing Classes:
CachePublisherImpl
@Experimental public interface CachePublisher<K,V>
Publisher to be used for non-blocking operations across the cache data set. Note that implementations must use the context of the Cache and the current thread it is invoked on. This means any current invocation context or Flags must be adhered to.

A CachePublisher is immutable and any method that returns a CachePublisher may return a new instance.

Note that any Function passed as an argument to any method on this interface will be marshalled and ran on the remote node. This means any state these Functions may reference will also be marshalled and these functions must not have any side effects.

The PublisherReducers and PublisherTransformers classes have utility methods that can be used as functions as necessary. Note that any arguments to these methods must also be marshallable still.

Note that any processor or data observed from the returned Publisher or Stage will be performed solely in the subscriber's node and thus can use data that cannot be marshalled and may have side effects as needed.

  • Method Details

    • parallelReduction

      CachePublisher<K,V> parallelReduction()
      Configures reduction methods to operate in parallel across nodes. This only affects methods like entryReduction(Function, Function) and keyReduction(Function, Function). The entryPublisher(Function) and keyPublisher(SerializableFunction) are not affected by this setting. Defaults to sequential.

      Care should be taken when setting this as it may cause adverse performance in some use cases.

      Returns:
      a publisher that will perform reduction methods in parallel

    • sequentialReduction

      CachePublisher<K,V> sequentialReduction()
      Configures reduction methods to operate in sequential across nodes. This only affects methods like entryReduction(Function, Function) and keyReduction(Function, Function). The entryPublisher(Function) and keyPublisher(SerializableFunction) are not affected by this setting. Defaults to sequential.
      Returns:
      a publisher that will perform reduction methods in sequential

    • batchSize

      CachePublisher<K,V> batchSize(int batchSize)
      Configures publisher methods chunk size when retrieving remote values. This only affects methods like entryPublisher(Function) and keyPublisher(Function). The entryReduction(Function, Function) and keyReduction(Function, Function) methods are unaffected. Defaults to the cache's configured StateTransferConfiguration.chunkSize().
      Parameters:
      batchSize - the size of the batch to use when retrieving remote entries
      Returns:
      a publisher that will perform publish methods with the given batch size
      Throws:
      IllegalArgumentException - if the provided value is 0 or less

    • withKeys

      CachePublisher<K,V> withKeys(Set<? extends K> keys)
      Configures the publisher to only publish keys or values that map to the given keys in the provided Set. Defaults to "all" keys, which can be done by invoking withAllKeys().

      Note that if this and withSegments(IntSet) are both used, then a key is only returned if it is also maps to a provided segment.

      Parameters:
      keys - set of keys that should only be used
      Returns:
      a publisher that will filter published values based on the given keys
      Throws:
      NullPointerException - if the provided Set is null

    • withAllKeys

      CachePublisher<K,V> withAllKeys()
      Configures a publisher to publish all keys or values, overriding if withKeys(Set) was invoked.
      Returns:
      a publisher that will return all keys or values.

    • withSegments

      CachePublisher<K,V> withSegments(IntSet segments)
      Configures the publisher to only publish keys or values that map to a given segment in the provided IntSet. The IntSets is recommended to be used. Defaults to "all" segments.

      Note that if this and withKeys(Set) are both used, then a key is only returned if it is also maps to a provided segment.

      Parameters:
      segments - determines what entries should be evaluated by only using ones that map to the given segments
      Returns:
      a publisher that will filter published values based on the given segments
      Throws:
      NullPointerException - if the provided IntSet is null

    • withAllSegments

      CachePublisher<K,V> withAllSegments()
      Configures the publisher to publish value(s) irrespective of their mapped segment. Defaults to "all" segments.
      Returns:
      a publisher that will publish all items irrespective of its segment

    • atMostOnce

      CachePublisher<K,V> atMostOnce()
      Allows the caller to configure what level of consistency is required when retrieving data. At most is the weakest guarantee which ensures that data in a segment will be read once in a stable topology, but if there is a concurrent topology update a given segment or a portion of its data may not be returned.

      The default data consistency is exactlyOnce().

    • atLeastOnce

      CachePublisher<K,V> atLeastOnce()
      Allows the caller to configure what level of consistency is required when retrieving data. At least ensures that data in a segment will be read once in a stable topology, but if there is a concurrent topology update a given segment or a portion of its data may be returned multiple times.

      The default data consistency is exactlyOnce().

      Returns:
      a publisher that will provide at least once data semantics

    • exactlyOnce

      CachePublisher<K,V> exactlyOnce()
      Allows the caller to configure what level of consistency is required when retrieving data. Exactly once ensures the highest level of guarantee so that even under a topology all data is returned once.

      Exactly once is the default data consistency level.

      Returns:
      a publisher that will provide exactly once data semantics

    • keyReduction

      <R> CompletionStage<R> keyReduction(Function<? super org.reactivestreams.Publisher<K>, ? extends CompletionStage<R>> transformer, Function<? super org.reactivestreams.Publisher<R>, ? extends CompletionStage<R>> finalizer)
      Same as entryReduction(Function, Function) except that the source publisher provided to the transformer is made up of keys only.
      Type Parameters:
      R - return value type
      Returns:
      CompletionStage that contains the resulting value when complete
      Throws:
      NullPointerException - if either the transformer or finalizer is null

    • keyReduction

      <R> CompletionStage<R> keyReduction(SerializableFunction<? super org.reactivestreams.Publisher<K>, ? extends CompletionStage<R>> transformer, SerializableFunction<? super org.reactivestreams.Publisher<R>, ? extends CompletionStage<R>> finalizer)
      Same as keyReduction(Function, Function) except that the Functions must also implement Serializable.

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

      Type Parameters:
      R - return value type
      Returns:
      CompletionStage that contains the resulting value when complete
      Throws:
      NullPointerException - if either the transformer or finalizer is null

    • entryReduction

      <R> CompletionStage<R> entryReduction(Function<? super org.reactivestreams.Publisher<CacheEntry<K,V>>, ? extends CompletionStage<R>> transformer, Function<? super org.reactivestreams.Publisher<R>, ? extends CompletionStage<R>> finalizer)
      Performs the given transformer and finalizer on data in the cache, resulting in a single value. Depending on the deliveryGuarantee the transformer may be invoked 1..numSegments times. It could be that the transformer is invoked for every segment and produces a result. All of these results are then fed into the finalizer to produce a final result. If publisher is parallel the finalizer will be invoked on each node to ensure there is only a single result per node.

      If the provided transformer internally uses a reduction with a default value, that value must be its identity value. This is the same as can be seen at Stream.reduce(Object, BinaryOperator). Then as long as the finalizer can handle the identity value it will be properly reduced.

      Type Parameters:
      R - return value type
      Parameters:
      transformer - reduces the given publisher of data eventually into a single value.
      finalizer - reduces all the single values produced by the transformer or this finalizer into one final value.
      Returns:
      CompletionStage that contains the resulting value when complete
      Throws:
      NullPointerException - if either the transformer or finalizer is null

    • entryReduction

      <R> CompletionStage<R> entryReduction(SerializableFunction<? super org.reactivestreams.Publisher<CacheEntry<K,V>>, ? extends CompletionStage<R>> transformer, SerializableFunction<? super org.reactivestreams.Publisher<R>, ? extends CompletionStage<R>> finalizer)
      Same as entryReduction(Function, Function) except that the Functions must also implement Serializable.

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

      Type Parameters:
      R - return value type
      Parameters:
      transformer - reduces the given publisher of data eventually into a single value.
      finalizer - reduces all the single values produced by the transformer or this finalizer into one final value.
      Returns:
      CompletionStage that contains the resulting value when complete
      Throws:
      NullPointerException - if either the transformer or finalizer is null

    • keyPublisher

      <R> SegmentPublisherSupplier<R> keyPublisher(Function<? super org.reactivestreams.Publisher<K>, ? extends org.reactivestreams.Publisher<R>> transformer)
      Same as entryPublisher(Function) except that the source publisher provided to the transformer is made up of keys only.
      Type Parameters:
      R - return value type
      Returns:
      Publisher that when subscribed to will return the results and notify of segment completion if necessary
      Throws:
      NullPointerException - if either the transformer is null

    • keyPublisher

      <R> SegmentPublisherSupplier<R> keyPublisher(SerializableFunction<? super org.reactivestreams.Publisher<K>, ? extends org.reactivestreams.Publisher<R>> transformer)
      Same as keyPublisher(Function) except that the Function must also implement Serializable.

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

      Type Parameters:
      R - return value type
      Returns:
      Publisher that when subscribed to will return the results and notify of segment completion if necessary
      Throws:
      NullPointerException - if either the transformer is null

    • entryPublisher

      <R> SegmentPublisherSupplier<R> entryPublisher(Function<? super org.reactivestreams.Publisher<CacheEntry<K,V>>, ? extends org.reactivestreams.Publisher<R>> transformer)
      Performs the given transformer on data in the cache, resulting in multiple values. If a single value is desired, the user should use entryReduction(Function, Function) instead as it can optimize some things. Depending on the deliveryGuarantee the transformer may be invoked 1..numSegments times per node. Results from a given node will retrieve values up to batchSize values until some are consumed.

      For example when using RxJava and using an intermediate operation such as Flowable.switchIfEmpty(Publisher) this can add elements if the given Publisher is empty, and it is very possible that a segment may not have entries and therefore may add the elements the switched Publisher returns multiple times.

      Methods that add elements to the returned Publisher are fine as long as they are tied to a specific entry, for example Flowable.flatMap(io.reactivex.rxjava3.functions.Function) which can reproduce the same elements when provided the same input entry from the cache.

      Type Parameters:
      R - return value type
      Parameters:
      transformer - transform the given stream of data into something else (requires non null)
      Returns:
      Publisher that when subscribed to will return the results and notify of segment completion if necessary
      Throws:
      NullPointerException - if either the transformer is null

    • entryPublisher

      <R> SegmentPublisherSupplier<R> entryPublisher(SerializableFunction<? super org.reactivestreams.Publisher<CacheEntry<K,V>>, ? extends org.reactivestreams.Publisher<R>> transformer)
      Same as entryPublisher(Function) except that the Function must also implement Serializable.

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

      Type Parameters:
      R - return value type
      Parameters:
      transformer - transform the given stream of data into something else (requires non null)
      Returns:
      Publisher that when subscribed to will return the results and notify of segment completion if necessary
      Throws:
      NullPointerException - if either the transformer is null