Search in sources :

Example 1 with StreamElementSerializer

use of org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer in project flink by apache.

the class AllWindowedStream method aggregate.

/**
	 * Applies the given window function to each window. The window function is called for each
	 * evaluation of the window for each key individually. The output of the window function is
	 * interpreted as a regular non-windowed stream.
	 *
	 * <p>Arriving data is incrementally aggregated using the given aggregate function. This means
	 * that the window function typically has only a single value to process when called.
	 *
	 * @param aggregateFunction The aggregation function that is used for incremental aggregation.
	 * @param windowFunction The window function.
	 * @param accumulatorType Type information for the internal accumulator type of the aggregation function
	 * @param resultType Type information for the result type of the window function
	 *
	 * @return The data stream that is the result of applying the window function to the window.
	 *
	 * @param <ACC> The type of the AggregateFunction's accumulator
	 * @param <V> The type of AggregateFunction's result, and the WindowFunction's input  
	 * @param <R> The type of the elements in the resulting stream, equal to the
	 *            WindowFunction's result type
	 */
@PublicEvolving
public <ACC, V, R> SingleOutputStreamOperator<R> aggregate(AggregateFunction<T, ACC, V> aggregateFunction, AllWindowFunction<V, R, W> windowFunction, TypeInformation<ACC> accumulatorType, TypeInformation<V> aggregateResultType, TypeInformation<R> resultType) {
    checkNotNull(aggregateFunction, "aggregateFunction");
    checkNotNull(windowFunction, "windowFunction");
    checkNotNull(accumulatorType, "accumulatorType");
    checkNotNull(aggregateResultType, "aggregateResultType");
    checkNotNull(resultType, "resultType");
    if (aggregateFunction instanceof RichFunction) {
        throw new UnsupportedOperationException("This aggregate function cannot be a RichFunction.");
    }
    //clean the closures
    windowFunction = input.getExecutionEnvironment().clean(windowFunction);
    aggregateFunction = input.getExecutionEnvironment().clean(aggregateFunction);
    final String callLocation = Utils.getCallLocationName();
    final String udfName = "AllWindowedStream." + callLocation;
    final String opName;
    final KeySelector<T, Byte> keySel = input.getKeySelector();
    OneInputStreamOperator<T, R> operator;
    if (evictor != null) {
        @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<T>> streamRecordSerializer = (TypeSerializer<StreamRecord<T>>) new StreamElementSerializer(input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        ListStateDescriptor<StreamRecord<T>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + evictor + ", " + udfName + ")";
        operator = new EvictingWindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalIterableAllWindowFunction<>(new AggregateApplyAllWindowFunction<>(aggregateFunction, windowFunction)), trigger, evictor, allowedLateness, lateDataOutputTag);
    } else {
        AggregatingStateDescriptor<T, ACC, V> stateDesc = new AggregatingStateDescriptor<>("window-contents", aggregateFunction, accumulatorType.createSerializer(getExecutionEnvironment().getConfig()));
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + udfName + ")";
        operator = new WindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalSingleValueAllWindowFunction<>(windowFunction), trigger, allowedLateness, lateDataOutputTag);
    }
    return input.transform(opName, resultType, operator).forceNonParallel();
}
Also used : StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) RichFunction(org.apache.flink.api.common.functions.RichFunction) AggregatingStateDescriptor(org.apache.flink.api.common.state.AggregatingStateDescriptor) ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) InternalSingleValueAllWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalSingleValueAllWindowFunction) InternalIterableAllWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalIterableAllWindowFunction) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer) PublicEvolving(org.apache.flink.annotation.PublicEvolving)

Example 2 with StreamElementSerializer

use of org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer in project flink by apache.

the class WindowedStream method apply.

/**
	 * Applies the given window function to each window. The window function is called for each
	 * evaluation of the window for each key individually. The output of the window function is
	 * interpreted as a regular non-windowed stream.
	 *
	 * <p>
	 * Arriving data is incrementally aggregated using the given fold function.
	 *
	 * @param initialValue The initial value of the fold.
	 * @param foldFunction The fold function that is used for incremental aggregation.
	 * @param function The window function.
	 * @param resultType Type information for the result type of the window function
	 * @return The data stream that is the result of applying the window function to the window.
	 *
	 * @deprecated Use {@link #fold(Object, FoldFunction, WindowFunction, TypeInformation, TypeInformation)} instead.
	 */
@Deprecated
public <R> SingleOutputStreamOperator<R> apply(R initialValue, FoldFunction<T, R> foldFunction, WindowFunction<R, R, K, W> function, TypeInformation<R> resultType) {
    if (foldFunction instanceof RichFunction) {
        throw new UnsupportedOperationException("FoldFunction of apply can not be a RichFunction.");
    }
    if (windowAssigner instanceof MergingWindowAssigner) {
        throw new UnsupportedOperationException("Fold cannot be used with a merging WindowAssigner.");
    }
    //clean the closures
    function = input.getExecutionEnvironment().clean(function);
    foldFunction = input.getExecutionEnvironment().clean(foldFunction);
    String callLocation = Utils.getCallLocationName();
    String udfName = "WindowedStream." + callLocation;
    String opName;
    KeySelector<T, K> keySel = input.getKeySelector();
    OneInputStreamOperator<T, R> operator;
    if (evictor != null) {
        @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<T>> streamRecordSerializer = (TypeSerializer<StreamRecord<T>>) new StreamElementSerializer(input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        ListStateDescriptor<StreamRecord<T>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + evictor + ", " + udfName + ")";
        operator = new EvictingWindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalIterableWindowFunction<>(new FoldApplyWindowFunction<>(initialValue, foldFunction, function, resultType)), trigger, evictor, allowedLateness, lateDataOutputTag);
    } else {
        FoldingStateDescriptor<T, R> stateDesc = new FoldingStateDescriptor<>("window-contents", initialValue, foldFunction, resultType.createSerializer(getExecutionEnvironment().getConfig()));
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + udfName + ")";
        operator = new WindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalSingleValueWindowFunction<>(function), trigger, allowedLateness, lateDataOutputTag);
    }
    return input.transform(opName, resultType, operator);
}
Also used : StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) RichFunction(org.apache.flink.api.common.functions.RichFunction) InternalSingleValueWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalSingleValueWindowFunction) ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) InternalIterableWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalIterableWindowFunction) FoldingStateDescriptor(org.apache.flink.api.common.state.FoldingStateDescriptor) MergingWindowAssigner(org.apache.flink.streaming.api.windowing.assigners.MergingWindowAssigner) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer)

Example 3 with StreamElementSerializer

use of org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer in project flink by apache.

the class WindowedStream method fold.

/**
	 * Applies the given window function to each window. The window function is called for each
	 * evaluation of the window for each key individually. The output of the window function is
	 * interpreted as a regular non-windowed stream.
	 *
	 * <p>Arriving data is incrementally aggregated using the given fold function.
	 *
	 * @param initialValue the initial value to be passed to the first invocation of the fold function
	 * @param foldFunction The fold function.
	 * @param foldResultType The result type of the fold function.
	 * @param windowFunction The process window function.
	 * @param windowResultType The process window function result type.
	 * @return The data stream that is the result of applying the fold function to the window.
	 */
@Internal
public <R, ACC> SingleOutputStreamOperator<R> fold(ACC initialValue, FoldFunction<T, ACC> foldFunction, ProcessWindowFunction<ACC, R, K, W> windowFunction, TypeInformation<ACC> foldResultType, TypeInformation<R> windowResultType) {
    if (foldFunction instanceof RichFunction) {
        throw new UnsupportedOperationException("FoldFunction can not be a RichFunction.");
    }
    if (windowAssigner instanceof MergingWindowAssigner) {
        throw new UnsupportedOperationException("Fold cannot be used with a merging WindowAssigner.");
    }
    //clean the closures
    windowFunction = input.getExecutionEnvironment().clean(windowFunction);
    foldFunction = input.getExecutionEnvironment().clean(foldFunction);
    String callLocation = Utils.getCallLocationName();
    String udfName = "WindowedStream." + callLocation;
    String opName;
    KeySelector<T, K> keySel = input.getKeySelector();
    OneInputStreamOperator<T, R> operator;
    if (evictor != null) {
        @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<T>> streamRecordSerializer = (TypeSerializer<StreamRecord<T>>) new StreamElementSerializer(input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        ListStateDescriptor<StreamRecord<T>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + evictor + ", " + udfName + ")";
        operator = new EvictingWindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalIterableProcessWindowFunction<>(new FoldApplyProcessWindowFunction<>(initialValue, foldFunction, windowFunction, foldResultType)), trigger, evictor, allowedLateness, lateDataOutputTag);
    } else {
        FoldingStateDescriptor<T, ACC> stateDesc = new FoldingStateDescriptor<>("window-contents", initialValue, foldFunction, foldResultType.createSerializer(getExecutionEnvironment().getConfig()));
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + udfName + ")";
        operator = new WindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalSingleValueProcessWindowFunction<>(windowFunction), trigger, allowedLateness, lateDataOutputTag);
    }
    return input.transform(opName, windowResultType, operator);
}
Also used : StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) RichFunction(org.apache.flink.api.common.functions.RichFunction) ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) FoldingStateDescriptor(org.apache.flink.api.common.state.FoldingStateDescriptor) MergingWindowAssigner(org.apache.flink.streaming.api.windowing.assigners.MergingWindowAssigner) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) InternalSingleValueProcessWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalSingleValueProcessWindowFunction) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer) InternalIterableProcessWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalIterableProcessWindowFunction) Internal(org.apache.flink.annotation.Internal)

Example 4 with StreamElementSerializer

use of org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer in project flink by apache.

the class WindowedStream method reduce.

/**
	 * Applies the given window function to each window. The window function is called for each
	 * evaluation of the window for each key individually. The output of the window function is
	 * interpreted as a regular non-windowed stream.
	 *
	 * <p>
	 * Arriving data is incrementally aggregated using the given reducer.
	 *
	 * @param reduceFunction The reduce function that is used for incremental aggregation.
	 * @param function The window function.
	 * @param resultType Type information for the result type of the window function.
	 * @param legacyWindowOpType When migrating from an older Flink version, this flag indicates
	 *                           the type of the previous operator whose state we inherit.
	 * @return The data stream that is the result of applying the window function to the window.
	 */
private <R> SingleOutputStreamOperator<R> reduce(ReduceFunction<T> reduceFunction, WindowFunction<T, R, K, W> function, TypeInformation<R> resultType, LegacyWindowOperatorType legacyWindowOpType) {
    if (reduceFunction instanceof RichFunction) {
        throw new UnsupportedOperationException("ReduceFunction of reduce can not be a RichFunction.");
    }
    //clean the closures
    function = input.getExecutionEnvironment().clean(function);
    reduceFunction = input.getExecutionEnvironment().clean(reduceFunction);
    String callLocation = Utils.getCallLocationName();
    String udfName = "WindowedStream." + callLocation;
    String opName;
    KeySelector<T, K> keySel = input.getKeySelector();
    OneInputStreamOperator<T, R> operator;
    if (evictor != null) {
        @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<T>> streamRecordSerializer = (TypeSerializer<StreamRecord<T>>) new StreamElementSerializer(input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        ListStateDescriptor<StreamRecord<T>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + evictor + ", " + udfName + ")";
        operator = new EvictingWindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalIterableWindowFunction<>(new ReduceApplyWindowFunction<>(reduceFunction, function)), trigger, evictor, allowedLateness, lateDataOutputTag);
    } else {
        ReducingStateDescriptor<T> stateDesc = new ReducingStateDescriptor<>("window-contents", reduceFunction, input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + udfName + ")";
        operator = new WindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, new InternalSingleValueWindowFunction<>(function), trigger, allowedLateness, lateDataOutputTag, legacyWindowOpType);
    }
    return input.transform(opName, resultType, operator);
}
Also used : ReducingStateDescriptor(org.apache.flink.api.common.state.ReducingStateDescriptor) StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) RichFunction(org.apache.flink.api.common.functions.RichFunction) InternalSingleValueWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalSingleValueWindowFunction) ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) InternalIterableWindowFunction(org.apache.flink.streaming.runtime.operators.windowing.functions.InternalIterableWindowFunction) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer)

Example 5 with StreamElementSerializer

use of org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer in project flink by apache.

the class WindowedStream method apply.

private <R> SingleOutputStreamOperator<R> apply(InternalWindowFunction<Iterable<T>, R, K, W> function, TypeInformation<R> resultType, String callLocation) {
    String udfName = "WindowedStream." + callLocation;
    SingleOutputStreamOperator<R> result = createFastTimeOperatorIfValid(function, resultType, udfName);
    if (result != null) {
        return result;
    }
    LegacyWindowOperatorType legacyWindowOpType = getLegacyWindowType(function);
    String opName;
    KeySelector<T, K> keySel = input.getKeySelector();
    WindowOperator<K, T, Iterable<T>, R, W> operator;
    if (evictor != null) {
        @SuppressWarnings({ "unchecked", "rawtypes" }) TypeSerializer<StreamRecord<T>> streamRecordSerializer = (TypeSerializer<StreamRecord<T>>) new StreamElementSerializer(input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        ListStateDescriptor<StreamRecord<T>> stateDesc = new ListStateDescriptor<>("window-contents", streamRecordSerializer);
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + evictor + ", " + udfName + ")";
        operator = new EvictingWindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, function, trigger, evictor, allowedLateness, lateDataOutputTag);
    } else {
        ListStateDescriptor<T> stateDesc = new ListStateDescriptor<>("window-contents", input.getType().createSerializer(getExecutionEnvironment().getConfig()));
        opName = "TriggerWindow(" + windowAssigner + ", " + stateDesc + ", " + trigger + ", " + udfName + ")";
        operator = new WindowOperator<>(windowAssigner, windowAssigner.getWindowSerializer(getExecutionEnvironment().getConfig()), keySel, input.getKeyType().createSerializer(getExecutionEnvironment().getConfig()), stateDesc, function, trigger, allowedLateness, lateDataOutputTag, legacyWindowOpType);
    }
    return input.transform(opName, resultType, operator);
}
Also used : StreamRecord(org.apache.flink.streaming.runtime.streamrecord.StreamRecord) ListStateDescriptor(org.apache.flink.api.common.state.ListStateDescriptor) TypeSerializer(org.apache.flink.api.common.typeutils.TypeSerializer) StreamElementSerializer(org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer)

Aggregations

StreamElementSerializer (org.apache.flink.streaming.runtime.streamrecord.StreamElementSerializer)33 ListStateDescriptor (org.apache.flink.api.common.state.ListStateDescriptor)30 TypeSerializer (org.apache.flink.api.common.typeutils.TypeSerializer)30 StreamRecord (org.apache.flink.streaming.runtime.streamrecord.StreamRecord)30 RichFunction (org.apache.flink.api.common.functions.RichFunction)17 Test (org.junit.Test)11 PublicEvolving (org.apache.flink.annotation.PublicEvolving)10 ConcurrentLinkedQueue (java.util.concurrent.ConcurrentLinkedQueue)9 AtomicInteger (java.util.concurrent.atomic.AtomicInteger)9 ExecutionConfig (org.apache.flink.api.common.ExecutionConfig)9 TypeHint (org.apache.flink.api.common.typeinfo.TypeHint)9 Tuple2 (org.apache.flink.api.java.tuple.Tuple2)9 KeyedOneInputStreamOperatorTestHarness (org.apache.flink.streaming.util.KeyedOneInputStreamOperatorTestHarness)9 FoldingStateDescriptor (org.apache.flink.api.common.state.FoldingStateDescriptor)6 ReducingStateDescriptor (org.apache.flink.api.common.state.ReducingStateDescriptor)6 MergingWindowAssigner (org.apache.flink.streaming.api.windowing.assigners.MergingWindowAssigner)6 GlobalWindow (org.apache.flink.streaming.api.windowing.windows.GlobalWindow)6 InternalIterableAllWindowFunction (org.apache.flink.streaming.runtime.operators.windowing.functions.InternalIterableAllWindowFunction)6 InternalSingleValueAllWindowFunction (org.apache.flink.streaming.runtime.operators.windowing.functions.InternalSingleValueAllWindowFunction)6 AggregatingStateDescriptor (org.apache.flink.api.common.state.AggregatingStateDescriptor)5