All Classes and Interfaces

Class
Description
A base implementation of AsyncStateSlicingWindowProcessor for window aggregate.
Abstract implementation for streaming unbounded Join operator based on async state api, which defines some member fields can be shared between different implementations.
Base class for TopN Function with async state api.
A base class for window aggregate processors with async state.
A checker that checks whether the window is empty.
Spilled files Merger of BinaryExternalSorter.
Bytes based hash map.
A binary map in the structure like Map<K, List<V>>, where there are multiple values under a single key, and they are all bytes based.
The AbstractMapBundleOperator simply used a java Map to store the input elements in key-value form.
AbstractRowDataSerializer<T extends org.apache.flink.table.data.RowData>
Row serializer, provided paged serialize paged method.
A basic implementation to support unbounded event-time over-window.
A base implementation of SlicingSyncStateWindowProcessor for window aggregate.
Abstract implementation for streaming unbounded Join operator which defines some member fields can be shared between different implementations.
Base class for TopN Function with sync state api.
A base class for window aggregate processors.
Base class for TopN Function.
An abstract helper to do the logic Top-n used for all top-n functions.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
The stream graph optimization strategy of adaptive broadcast join.
Interface for implementing an adaptive join operator.
Adaptive join factory.
Utils for adaptive join optimization strategies.
The stream graph optimization strategy of adaptive skewed join.
An implementation of RecordsCombiner that accumulates input records into the window accumulator state.
Factory to create AggCombiner.
A WindowOperator for grouped window aggregates.
The base class for handling aggregate functions.
The base class for handling aggregate or table aggregate functions.
The operator for aligned window table function.
A variant of AppendOnlyTopNFunction to handle first-n case.
A TopN function could handle insert-only stream.
A helper to help do the logic 'Top-n' for append-only stream in AppendOnlyTopNFunction and AsyncStateAppendOnlyTopNFunction.
Built-in ARRAY_AGG aggregate function.
Accumulator for ARRAY_AGG with retraction.
Implementation of BuiltInFunctionDefinitions.ARRAY_APPEND.
Converter for ArrayType of boolean[] external type.
Converter for ArrayType of byte[] external type.
Implementation of BuiltInFunctionDefinitions.ARRAY_CONCAT.
Implementation of BuiltInFunctionDefinitions.ARRAY_CONTAINS.
Serializer for ArrayData.
TypeSerializerSnapshot for ArrayDataSerializer.
Implementation of BuiltInFunctionDefinitions.ARRAY_DISTINCT.
Converter for ArrayType of double[] external type.
Implementation of BuiltInFunctionDefinitions.ARRAY_EXCEPT.
Converter for ArrayType of float[] external type.
Converter for ArrayType of int[] external type.
Implementation of BuiltInFunctionDefinitions.ARRAY_INTERSECT.
Implementation of BuiltInFunctionDefinitions.ARRAY_JOIN.
Converter for ArrayType of List external type.
Converter for ArrayType of long[] external type.
Implementation of BuiltInFunctionDefinitions.ARRAY_MAX.
Implementation of BuiltInFunctionDefinitions.ARRAY_MIN.
Converter for ArrayType of nested primitive or object arrays external types.
Implementation of BuiltInFunctionDefinitions.ARRAY_POSITION.
Implementation of BuiltInFunctionDefinitions.ARRAY_PREPEND.
Implementation of BuiltInFunctionDefinitions.ARRAY_REMOVE.
Implementation of BuiltInFunctionDefinitions.ARRAY_REVERSE.
Converter for ArrayType of short[] external type.
Implementation of BuiltInFunctionDefinitions.ARRAY_SLICE.
Implementation of ARRAY_SORT function.
Implementation of BuiltInFunctionDefinitions.ARRAY_UNION.
The AssociatedRecords is the records associated to the input row.
Async function runner for AsyncScalarFunction, which takes the generated function, instantiates it, and then calls its lifecycle methods.
The async join runner to lookup the dimension table.
The async join runner with an additional calculate function on the dimension table.
Callback to be used in when merging slices or windows for specifying which slices or windows should be merged.
An implementation of RecordsCombiner that accumulates input records into the window accumulator with async state.
Factory to create AsyncStateAggCombiner.
A TopN function could handle insert-only stream.
A more concise implementation for AppendOnlyTopNFunction and UpdatableTopNFunction when only Top-1 is desired.
Aggregate Function used for the groupby (without window) aggregate with async state api.
Context to switch current key in async state backend.
The AsyncStateRecordsCombiner is used to combine buffered records into async state.
A factory that creates a AsyncStateRecordsCombiner.
An implementation of AsyncStateWindowBuffer that buffers input elements in a WindowBytesMultiMap and combines buffered elements into async state when flushing.
This function is used to deduplicate on keys and keeps only first or last row on row time.
A window aggregate processor implementation which works for SliceSharedAssigner with async state api, e.g. hopping windows and cumulative windows.
A window aggregate processor implementation which works for SliceUnsharedAssigner with async state, e.g. tumbling windows.
The AsyncStateSlicingWindowProcessor is an optimized processing for aligned windows which can apply the slicing optimization.
Streaming unbounded Join operator based on async state api, which supports INNER/LEFT/RIGHT/FULL JOIN.
Table operator to invoke close always.
Utils for processing with async state.
A processor that processes elements for windows with async state api.
A buffer that buffers data in memory and flushes many values to async state together at a time to avoid frequently accessing async state, or flushes to output to reduce shuffling data.
A factory that creates a WindowBuffer with async state.
A AsyncStateWindowJoinOperator implemented by async state api.
A processor that processes elements for windows with async state.
Information available in an invocation of methods of AsyncStateWindowProcessor.
The base stream graph optimization strategy class for adaptive join operator.
An abstract TwoInputStreamOperator that allows its subclasses to clean up their state based on a TTL.
Built-in APPROX_COUNT_DISTINCT aggregate function for Batch sql.
Base function for APPROX_COUNT_DISTINCT aggregate.
Built-in byte APPROX_COUNT_DISTINCT aggregate function.
Built-in Date APPROX_COUNT_DISTINCT aggregate function.
Built-in decimal APPROX_COUNT_DISTINCT aggregate function.
Built-in double APPROX_COUNT_DISTINCT aggregate function.
Built-in float APPROX_COUNT_DISTINCT aggregate function.
Built-in int APPROX_COUNT_DISTINCT aggregate function.
Built-in long APPROX_COUNT_DISTINCT aggregate function.
Built-in Short APPROX_COUNT_DISTINCT aggregate function.
Built-in string APPROX_COUNT_DISTINCT aggregate function.
Built-in Time APPROX_COUNT_DISTINCT aggregate function.
Built-in Timestamp APPROX_COUNT_DISTINCT aggregate function.
Built-in TimestampLtz APPROX_COUNT_DISTINCT aggregate function.
A MultipleInputStreamOperatorBase to handle batch operators.
The factory to create BatchMultipleInputStreamOperator.
TypeInformation for BigDecimal.
Writer for binary array.
Accessor for setting the elements of an array writer to null during runtime.
Record merger for sort of BinaryRowData.
The BinaryExternalSorter is a full fledged sorter for binary format.
Bucket area for hash table.
A partition.
Build side buffer.
Hash partitioner for BinaryRowData.
An implementation of a Hybrid Hash Join.
abstract sortable, provide basic compare and swap.
In memory sort buffer for binary row.
Key-Value style record merger for sort.
In memory KV sortable buffer for binary row, it already has records in memory.
Binary version of MergeIterator.
A simple iterator over the input read though an I/O channel.
A KeySelector which will extract key from RowData.
Serializer for BinaryRowData.
TypeSerializerSnapshot for BinaryRowDataSerializer.
Utilities for BinaryRowData.
Writer for BinaryRowData.
Util for BinaryStringData.
Writer to write a composite data format, like row, array. 1.
Accessor for setting the elements of an array writer during runtime.
A watermark generator for rowtime attributes which are out-of-order by a bounded time interval.
An implementation of RowData which also is also backed by an array of Java Object, just similar to GenericRowData.
An Output that can be used to emit elements and other messages to multiple outputs.
Implementation of BuiltInFunctionDefinitions.BTRIM.
A BufferBundle is a bundle to buffer the input records in memory and fold data based on specified pattern to reduce state access.
the operator for OVER window need cache data by ResettableExternalBuffer for OverWindowFrame.
Sorter for buffered input in the form of Key-Value Style.
Iterate all the elements in memory which has not been(or has been) probed during probe phase.
Partition bucket iterator.
Base class for runtime implementation represented as AggregateFunction that is constructed from BuiltInFunctionDefinition.specialize(SpecializedContext).
Base class for runtime implementation represented as ScalarFunction that is constructed from BuiltInFunctionDefinition.specialize(SpecializedContext).
Base class for built-in functions that need another level of specialization via BuiltInFunctionDefinition.specialize(SpecializedContext).
Base class for runtime implementation represented as TableFunction that is constructed from BuiltInFunctionDefinition.specialize(SpecializedContext).
A BundleTrigger determines when a bundle of input elements should be evaluated and trigger the callback which registered previously.
Interface for bundle trigger callbacks that can be registered to a BundleTrigger.
Byte hash set.
A binary map in the structure like Map<BinaryRowData, BinaryRowData>.
MemorySegmentPool for BytesHashMap to fall back to sort agg.
Base class for BytesHashMap and BytesMultiMap.
Result fetched when looking up a key.
A binary map in the structure like Map<BinaryRowData, List<BinaryRowData>>.
Abstract task that loads data in Full cache from source provided by ScanTableSource.ScanRuntimeProvider.
A wrapper function around user-provided async lookup function with a cache layer.
A wrapper function around user-provided lookup function with a cache layer.
Key-Value style channel reader input view iterator.
Channel with block count and numBytesInLastBlock of file.
Utils for get DataType from a Class.
Get internal(sql engine execution data formats) and default external class for LogicalType.
Base interface for clean up state, both for ProcessFunction and CoProcessFunction.
A clock service which can get current processing time.
Implementation of BuiltInFunctionDefinitions.COALESCE.
A CoBundleTrigger is similar with BundleTrigger, and the only differences is CoBundleTrigger can handle two inputs.
Stream operator factory for code gen operator.
Built-in COLLECT aggregate function.
Accumulator for COLLECT.
Because it's impossible to restore a RecordComparator instance generated by GeneratedRecordComparator from checkpoint snapshot.
Utilities to compile a generated code to a Class.
Compressed block channel reader provides a scenario where MemorySegment must be maintained.
Compressed block channel writer provides a scenario where MemorySegment must be maintained.
A DataInputView that is backed by a BufferFileReader, making it effectively a data input stream.
A DataOutputView that is backed by a FileIOChannel, making it effectively a data output stream.
rankStart and rankEnd are inclusive, rankStart always start from one.
ConstantRankRangeWithoutEnd is a RankRange which not specify RankEnd.
Processes RowData to enforce the following constraints: NOT NULL column constraint of a sink table CHAR(length)/@{code VARCHAR(length)}: trim string values to comply with the length defined in their corresponding types.
Helper builder, so that the ConstraintEnforcer can be instantiated with only the NOT NULL constraint validation, only the CHAR/VARCHAR length validation, only the BINARY/VARBINARY length validation or combinations of them, or all of them.
Helper POJO to keep info about CHAR/VARCHAR/BINARY/VARBINARY fields, used to determine if trimming or padding is needed.
Special version of BroadcastingOutput that performs a shallow copy of the StreamRecord to ensure that multi-output works correctly.
An Output that can be used to emit copying elements and other messages for the first input of TwoInputStreamOperator.
An Output that can be used to emit copying elements and other messages for OneInputStreamOperator.
An Output that can be used to emit copying elements and other messages for the second input of TwoInputStreamOperator.
A BundleTrigger that fires once the count of elements in a bundle reaches the given count.
A CoBundleTrigger that fires once the count of elements in a bundle reaches the given count.
A GroupWindowAssigner that windows elements into sliding windows based on the count number of the elements.
A GroupWindowAssigner that windows elements into fixed-size windows based on the count number of the elements.
A Window that represents a count window.
The serializer used to write the CountWindow type.
Serializer configuration snapshot for compatibility and format evolution.
A GroupWindowAssigner that windows elements into cumulative windows based on the timestamp of the elements.
Converters between internal data format and java format.
Abstract converter for internal base row.
Converter for ArrayData.
Converter for BigDecimal.
Converter for boolean.
Converter for byte.
Converter for case class.
Converter between internal data format and java format.
Converter for date.
Converter for DecimalData.
Converter for double.
Converter for float.
Converter for RawValueData.
Identity converter.
Converter for Instant.
Converter for int.
Converter for LocalDate.
Converter for LocalDateTime.
Converter for LocalTime.
Converter for long.
Converter for Long and TimestampData.
Converter for map.
Converter for MapData.
Converter for object array.
Converter for pojo.
Converter for primitive boolean array.
Converter for primitive byte array.
Converter for primitive double array.
Converter for primitive float array.
Converter for primitive int array.
Converter for primitive long array.
Converter for primitive short array.
Converter for RawValueData.
Converter for row.
Converter for base row.
Converter for short.
Converter for String.
Converter for StringData.
Converter for time.
Converter for timestamp.
Converter for TimestampData class.
Converter for timestamp which doesn't consider the time zone.
Converter for flink tuple.
Converter between internal and external data structure.
Registry of available data structure converters.
The data type visitor used to fix the precision for data type with the given logical type which carries the correct precisions.
Information about a DataView stored in state.
Converter for DateType of Date external type.
Converter for DateType of LocalDate external type.
Converter for DayTimeIntervalType of Duration external type.
Converter for DecimalType of BigDecimal external type.
Serializer for DecimalData.
TypeSerializerSnapshot for DecimalDataSerializer.
TypeInformation for DecimalData.
Utilities for DecimalData.
Base class for deduplicate function.
Utility for deduplicate function.
Default runtime implementation for SpecializedFunction.ExpressionEvaluator.
Inspired by DelegatingResultFuture for AsyncScalarFunction.
Delegates actions of CompletableFuture to ResultFuture.
Double hash set.
A function drops only rows with RowKind.UPDATE_BEFORE changelog kind.
Operator to collect and build the DynamicFilteringData for sources that supports dynamic filtering.
The operator coordinator for DynamicFilteringDataCollectorOperator.
A Trigger that fires at some point after a specified number of input elements have arrived.
A Trigger that fires at some point after a specified number of input elements have arrived.
A Trigger that triggers on every element.
Implementation of BuiltInFunctionDefinitions.ELT.
A utility class which key is always empty no matter what the input row is.
Implementation of BuiltInFunctionDefinitions.ENDS_WITH.
This class is used for scalar function, in that it is used for sharing the initialization context between scalar functions that need code generated hashcode and equals method.
A Trigger that reacts to event-time timers.
A Trigger that fires once the watermark passes the end of the window to which a pane belongs.
A composite Trigger that consist of AfterEndOfWindow and a early trigger and late trigger.
A composite Trigger that consist of AfterEndOfWindow and a late trigger.
A special exception that signifies that the cause exception came from a MultipleInputStreamOperatorBase.
A ExecutionContext contains information about the context in which functions are executed and the APIs to create state.
Implementation of ExecutionContext.
A serializer that can serialize and deserialize all data structures defined by a DataType.
TypeSerializerSnapshot for ExternalSerializer.
Type information that wraps a serializer that originated from a DataType.
A more concise implementation for AppendOnlyTopNFunction and UpdatableTopNFunction when only Top-1 is desired.
A helper to help do the logic 'Top-1' in FastTop1Function and AsyncStateFastTop1Function.
File channel util for runtime.
Function filtering out all the input records.
Interface for code generated filter condition function on single RowData.
Input for the first input of TwoInputStreamOperator.
An Output that can be used to emit elements and other messages for the first input of TwoInputStreamOperator.
Built-in FIRST_VALUE aggregate function.
Built-in FIRST_VALUE with retraction aggregate function.
Accumulator for FIRST_VALUE.
Join type for join.
Float hash set.
Base MultipleInputStreamOperator to handle multiple operator fusion codegen in table module.
The general implementation of InternalWindowProcessFunction.
Describes a generated AggsHandleFunction.
A wrapper for generated class, defines a GeneratedClass.newInstance(ClassLoader) method to get an instance by reference objects easily.
GeneratedCollector<C extends org.apache.flink.util.Collector<?>>
Describes a generated Collector.
Describes a generated FilterCondition.
GeneratedFunction<F extends org.apache.flink.api.common.functions.Function>
Describes a generated Function.
Describes a generated HashFunction.
GeneratedInput<F extends org.apache.flink.api.common.io.InputFormat<?,?>>
Describes a generated InputFormat.
Describes a generated JoinCondition.
Describes a generated NamespaceAggsHandleFunction.
Describes a generated NamespaceTableAggsHandleFunction.
Describes a generated NormalizedKeyComputer.
GeneratedOperator<C extends org.apache.flink.streaming.api.operators.StreamOperator<?>>
Describes a generated StreamOperator.
Describes a generated Projection.
Describes a generated RecordComparator.
Describes a generated RecordEqualiser.
GeneratedResultFuture<T extends org.apache.flink.streaming.api.functions.async.ResultFuture<?>>
Describes a generated ResultFuture.
Describes a generated TableAggsHandleFunction.
Describes a generated WatermarkGenerator.
Wrapper of the GeneratedWatermarkGenerator that is used to create WatermarkGenerator.
Wrapper of the code-generated WatermarkGenerator.
A KeySelector which will extract key from RowData.
An implementation of RecordsCombiner that accumulates local accumulators records into the window accumulator state.
Factory to create GlobalAggCombiner.
Global runtime filter builder operator.
Aggregate Function used for the groupby (without window) aggregate.
A helper to do the logic of group agg.
Aggregate Function used for the groupby (without window) table aggregate.
A GroupWindowAssigner assigns zero or more Windows to an element.
Interface for code generated hash code of RowData which will select some fields to hash or ArrayData or MapData.
Hash join base operator.
Join type for hash table.
A DataOutputView that is backed by a BlockChannelWriter, making it effectively a data output stream.
A jvm heap implementation of WindowsGrouping, which uses a linked list to buffer all the inputs of a keyed group belonging to the same window.
Buffer used by HyperLogLogPlusPlus and ApproximateCountDistinct.
The implement of HyperLogLogPlusPlus is inspired from Apache Spark.
No-op converter that just forwards its input.
Implementation of BuiltInFunctionDefinitions.IF_NULL.
Operator that converts to internal data structures and wraps atomic records if necessary.
CacheLoader that used InputFormat for loading data into the cache.
This handler is mainly used for selecting the next available input index according to read priority in MultipleInputStreamOperatorBase.
Describe the inputId and read order of MultipleInput operator.
For the case that input has no uniqueKey.
For the case that input has uniqueKey which is not contained by joinKey.
Describe the info of Input.
Parallel task that loads data into the cache from InputFormat with specified InputSplit.
The insensitive window frame calculates the statements which shouldn't care the window frame, for example RANK/DENSE_RANK/PERCENT_RANK/NTILE/ROW_NUMBER.
TypeSerializer of LogicalType for internal data structures.
The internal time window assigner which has some useful methods.
Type information that wraps a serializer that originated from a LogicalType.
The internal interface for functions that process over grouped windows.
Information available in an invocation of methods of InternalWindowProcessFunction.
IntervalJoinFunction is a RichFlatJoinFunction wrappers a GeneratedJoinCondition.
Int hash set.
A RichIterativeCondition wrapper to delegate invocation to the code generated RichIterativeCondition.
Interface for code generated condition function for [[org.apache.calcite.rel.core.Join]].
Utility to take null filters into consideration when apply join condition.
A helper to do the logic of streaming join.
The JoinInputSideSpec is ap specification which describes input side information of a Join.
For the case that input has joinKey contains uniqueKey.
A JoinRecordAsyncStateView is a view to the join state.
Utility to create a JoinRecordAsyncStateView depends on JoinInputSideSpec.
A JoinRecordStateView is a view to the join state.
Utility to create a JoinRecordStateView depends on JoinInputSideSpec.
Implementation for BuiltInFunctionDefinitions.JSON_ARRAYAGG_ABSENT_ON_NULL / BuiltInFunctionDefinitions.JSON_ARRAYAGG_NULL_ON_NULL.
Accumulator for JsonArrayAggFunction.
Implementation for BuiltInFunctionDefinitions.JSON_OBJECTAGG_NULL_ON_NULL / BuiltInFunctionDefinitions.JSON_OBJECTAGG_ABSENT_ON_NULL.
Accumulator for JsonObjectAggFunction.
The default cache for the jsonpath com.jayway.jsonpath.spi.cache.CacheProvider.
Implementation of BuiltInFunctionDefinitions.JSON_QUOTE.
Implementation of BuiltInFunctionDefinitions.JSON_UNQUOTE.
A KeyedCoProcessOperator that supports holding back watermarks with a static delay.
The lookup join handler which holds a LookupJoinRunner to process lookup for insert or update_after record and directly process delete and update_before record via local state.
The KeyedMapBundleOperator uses framework's key as bundle map key, thus can only be used on KeyedStream.
A function that processes elements of a stream, and could cleanup state.
An internal iterator interface which presents a more restrictive API than Iterator.
Lag AggregateFunction.
Accumulator for LAG.
Built-in LAST_VALUE aggregate function.
Built-in LAST_VALUE with retraction aggregate function.
Accumulator for LAST_VALUE with retraction.
MemorySegmentPool that lazy allocate memory pages from MemoryManager.
TypeInformation for Instant.
TypeInformation for LocalDateTime.
TypeInformation for Timestamp.
Operator for batch limit.
A serializer for LinkedList.
Snapshot class for the LinkedListSerializer.
Built-in LISTAGG with retraction aggregate function.
Accumulator for LISTAGG with retraction.
Built-in LISTAGGWS with retraction aggregate function.
Accumulator for LISTAGGWS with retraction.
A listenable collector for lookup join that can be called when an original record was collected.
An interface can listen on collecting original record.
MemorySegment pool of a MemorySegment list.
Specification for a ListView.
An implementation of RecordsCombiner that accumulates input records into local accumulators.
Factory to create LocalAggCombiner.
Local runtime filter builder operator.
The operator used for local window aggregation.
Converter for LocalZonedTimestampType of Instant external type.
Converter for LocalZonedTimestampType of Integer external type.
Converter for LocalZonedTimestampType of Long external type.
Converter for LocalZonedTimestampType of Timestamp external type.
Deprecated.
Partition for LongHybridHashTable.
A simple iterator over the input read though an I/O channel.
Long hash set.
Special optimized hashTable with key long.
Build iterator from bucket to match probe row.
Managing shared caches across different subtasks.
A wrapper class of LookupCache which also tracks the reference count of it.
Internal implementation of LookupCache for LookupOptions.LookupCacheType.FULL.
The join runner to lookup the dimension table.
The join runner with an additional calculate function on the dimension table.
Implementation of DynamicTableSource.Context.
An LRU cache, based on LinkedHashMap.
An object that can receive a notification when an entry is removed from a LRUMap.
Basic interface for map bundle processing.
The MapBundleOperator uses a KeySelector to extract bundle key, thus can be used with non-keyed-stream.
Serializer for MapData.
TypeSerializerSnapshot for ArrayDataSerializer.
Utilities for MapData.
Implementation of BuiltInFunctionDefinitions.MAP_ENTRIES.
Implementation of BuiltInFunctionDefinitions.MAP_FROM_ARRAYS.
Implementation of BuiltInFunctionDefinitions.MAP_KEYS.
Converter for MapType/MultisetType of Map external type.
Implementation of BuiltInFunctionDefinitions.MAP_UNION.
Implementation of BuiltInFunctionDefinitions.MAP_VALUES.
Specification for a MapView.
Built-in MAX with retraction aggregate function.
Accumulator for MAX with retraction.
MemorySegment pool to hold pages in memory.
Callback to be used in when merging slices or windows for specifying which slices or windows should be merged.
A WindowAssigner that can merge windows.
A default merging consumer that merges the accumulators in state windows that waiting to be merged into the target state window.
Utility for keeping track of merging Windows when using a MergingWindowAssigner in a WindowOperator.
Aggregate Function used for the global groupby (without window) aggregate in miniBatch mode.
Aggregate Function used for the groupby (without window) aggregate in miniBatch mode.
Aggregate Function used for the incremental groupby (without window) aggregate in miniBatch mode.
Aggregate Function used for the local groupby (without window) aggregate in miniBatch mode.
Streaming unbounded Join base operator which support mini-batch join.
Built-in MIN with retraction aggregate function.
Accumulator for MIN with retraction.
Base MultipleInputStreamOperator to handle multiple inputs in table module.
Murmur hasher.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
The base class for handling aggregate functions with namespace.
The base class for handling aggregate or table aggregate functions.
The base class for handling table aggregate functions with namespace.
The operator for OVER window don't need cache data.
Normalized key computer for BinaryInMemorySortBuffer.
Helper for null aware join.
An Iterator implementation that iterate on a map iterator and a null entry.
A Map Entry that the entry key is always null.
This class is used for scalar function, in that it is used for comparing Objects using code generated hashCode and equals instead of using the Object#equals/Object#hashcode versions.
Wrap HashSet with hashSet interface.
The offset window frame calculates frames containing LEAD/LAG statements.
Calc offset from base row.
Input for OneInputStreamOperator.
An Output that can be used to emit elements and other messages for OneInputStreamOperator.
Stream operator factory for multiple operator fusion codegen.
A type-specific hash set with with a fast, small-footprint implementation.
An utility to generate reusable padding results for outer joins.
Utility to create a OuterJoinRecordAsyncStateViews depends on JoinInputSideSpec.
Utility to create a OuterJoinRecordStateViews depends on JoinInputSideSpec.
An OuterRecord is a composite of record and numOfAssociations.
Base Output that can be used to emit elements and other messages in MultipleInputStreamOperatorBase.
Operator that converts to external data structures and unwraps atomic records if necessary.
A window frame calculates the results for those records belong to a window frame.
Function performing left padding.
Function performing right padding.
A type serializer which provides paged serialize and deserialize methods.
A WindowAssigner that window can be split into panes.
A PatternProcessFunction wrapper to delegate invocation to the code generated PatternProcessFunction.
Built-in PERCENTILE aggregate function.
Percentile agg function with multiple percentage parameters.
Accumulator for PERCENTILE.
Percentile agg function with only one percentage parameter.
Generates periodic watermarks based on a PeriodicWatermarkAssigner.
Default implementation of StateDataViewStore that currently forwards state registration to a RuntimeContext.
An implementation of StateDataViewStore for window aggregates which forwards the state registration to an underlying KeyedStateBackend.
Utilities for LogicalType and DataType..
The post-processing phase of adaptive join optimization, which must be placed at the end of all adaptive join optimization strategies.
Implementation of BuiltInFunctionDefinitions.PRINTF.
Probe iterator from probe or spilled partition.
A Trigger that reacts to processing-time timers.
A Trigger that fires once the current system time passes the end of the window to which a pane belongs.
A composite Trigger that consist of AfterEndOfWindow and a early trigger.
Trigger every a given interval, the first trigger time is interval after the first element in the pane.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
This function is used to deduplicate on keys and keeps only first row.
This function is used to deduplicate on keys and keeps only last row.
The function to execute processing time interval stream inner-join.
A stream operator that emits mini-batch marker in a given period.
This function is used to get the first row for every key partition in miniBatch mode.
This function is used to get the last row for every key partition in miniBatch mode.
Process Function used for the aggregate in bounded proc-time OVER window.
Process Function for ROW clause processing-time bounded OVER window.
Sort on proc-time and additional secondary sort attributes.
Process Function for processing-time unbounded OVER window.
Projection<IN extends org.apache.flink.table.data.RowData,OUT extends org.apache.flink.table.data.RowData>
Interface for code generated projection, which will map a RowData to another one.
Generates periodic watermarks based on a PunctuatedWatermarkAssigner.
The range sliding window frame calculates frames with the following SQL form: ...
The range unboundedFollowing window frame calculates frames with the following SQL form: ...
The range UnboundPreceding window frame calculates frames with the following SQL form: ...
Rank operator to compute top N.
RankRange for Rank, including following 3 types : ConstantRankRange, ConstantRankRangeWithoutEnd, VariableRankRange.
An enumeration of rank type, usable to show how exactly generate rank number.
Converter for RawType of byte[] external type.
Deserialization schema from raw (byte based) value to Flink Table/SQL internal data structure RowData.
Format factory for raw format which allows to read and write raw (byte based) values as a single column.
Options for the "raw" format.
Serialization schema that serializes an RowData object into raw (byte based) value.
Converter for RawType of object external type.
Serializer for RawValueData.
TypeSerializerSnapshot for RawValueDataSerializer.
Record comparator for BinaryInMemorySortBuffer.
The RecordCounter is used to count the number of input records under the current key.
Record equaliser for RowData which can compare two RowData and returns whether they are equal.
The RecordsCombiner is used to combine buffered records into state.
A factory that creates a RecordsCombiner.
A factory that creates a RecordsCombiner used for combining at local stage.
An implementation of WindowBuffer that buffers input elements in a WindowBytesMultiMap and combines buffered elements into state when flushing.
Implementation of BuiltInFunctionDefinitions.REGEXP_COUNT.
Implementation of BuiltInFunctionDefinitions.REGEXP_EXTRACT_ALL.
Implementation of BuiltInFunctionDefinitions.REGEXP_INSTR.
Implementation of BuiltInFunctionDefinitions.REGEXP_SUBSTR.
Runtime implementation of CacheReloadTrigger.Context.
Replicate the row N times.
A resettable external buffer for binary row.
Resettable buffer that add RowData and return BinaryRowData iterator.
Row iterator that can be reset.
A utility class to wrap the data stream api AsyncRetryStrategy to support both sync and async retry in table module.
A TopN function could handle updating stream.
A delegator holds user's AsyncLookupFunction to handle retries.
A delegator holds user's LookupFunction to handle retries.
Contains retry predicates used to determine if a result or error should result in a retry.
Returns true for any exception.
Returns true if the response is null or empty.
The partitioner used to partition row data by the connector's custom logic.
An implementation of EventComparator based on a generated RecordComparator.
RowDataKeySelector takes an RowData and extracts the deterministic key for the RowData.
Serializer for RowData.
TypeSerializerSnapshot for BinaryRowDataSerializer.
Utilities for RowData.
RowIterator<T extends org.apache.flink.table.data.RowData>
An internal iterator interface which presents a more restrictive API than Iterator.
An operator that sets the row kind of the incoming records to a specific row kind.
Converter for RowType of Row external type.
The row sliding window frame calculates frames with the following SQL form: ...
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
This function is used to deduplicate on keys and keeps only first or last row on row time.
A helper to deduplicate data with row time in RowTimeDeduplicateFunction and AsyncStateRowTimeDeduplicateFunction.
An implementation of RecordsCombiner that stores the first/last records of incremental input records into the window state.
The function to execute row(event) time interval stream inner-join.
A stream operator that emits watermark in a given event-time interval.
This function is used to get the first or last row for every key partition in miniBatch mode.
This function is used to get the first or last row for every key partition in miniBatch mode.
Process Function for RANGE clause event-time bounded OVER window.
A ProcessFunction to support unbounded RANGE window.
Process Function for ROWS clause event-time bounded OVER window.
A ProcessFunction to support unbounded ROWS window.
Sort based on event-time and possibly additional secondary sort attributes.
A rowtime window deduplicate processor.
The RowTimeWindowDeduplicateOperatorBuilder is used to build a WindowAggOperator for rowtime window deduplicate.
The row unboundedFollowing window frame calculates frames with the following SQL form: ...
The row UnboundPreceding window frame calculates frames with the following SQL form: ...
Utilities for runtime filter.
Implementation of DynamicTableSource.Context.
The wrapper class of the input/out schema for script transform.
Input for the second input of TwoInputStreamOperator.
An Output that can be used to emit elements and other messages for the second input of TwoInputStreamOperator.
Util for data format segments calc.
A GroupWindowAssigner that windows elements into sessions based on the timestamp.
Short hash set.
A Simple gauge providing method for value updating.
Copied from SingleElementIterator.
A StreamOperator for executing SinkFunctions.
Implementation of DynamicTableSink.Context.
An operator that maintains incoming records in state corresponding to the upsert keys and generates an upsert view for the downstream operator.
A SliceAssigner assigns element into a single slice.
Utilities to create SliceAssigners.
The SliceAssigner for cumulative windows.
The SliceAssigner for hopping windows.
The SliceAssigner for elements have been attached slice end timestamp, and the slices are shared.
The SliceAssigner for elements have been attached slice end timestamp, but the slices are not shared, i.e. the assigned slice is equal to the final window.
The SliceAssigner for tumbling windows.
The SliceAssigner for elements have been attached window start and end timestamps.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
A SliceAssigner which shares slices for windows, which means a window is divided into multiple slices and need to merge the slices into windows when emitting windows.
A window aggregate processor implementation which works for SliceSharedAssigner, e.g.
A SliceAssigner which doesn't share slices for windows, which means a window is divided into only one slice and doesn't need to merge the slices when emitting windows.
A window aggregate processor implementation which works for SliceUnsharedAssigner, e.g.
The SlicingSyncStateWindowProcessor is an optimized processing for aligned windows which can apply the slicing optimization.
A WindowTimerService for slicing window.
The sliding window frame.
A GroupWindowAssigner that windows elements into sliding windows based on the timestamp of the elements.
A serializer for SortedMap.
Snapshot class for the SortedMapSerializer.
The type information for sorted maps.
Operator for batch sort limit.
Gets two matched rows for full outer join.
Gets probeRow and match rows for inner join.
This function is used to process the main logic of sort merge join.
Gets probeRow and match rows for inner/left/right join.
An implementation that realizes the joining through a sort-merge join strategy. 1.In most cases, its performance is weaker than HashJoin. 2.It is more stable than HashJoin, and most of the data can be sorted stably. 3.SortMergeJoin should be the best choice if sort can be omitted in the case of multi-level join cascade with the same key.
Gets probeRow and match rows for left/right join.
Operator for batch sort.
Util for sort.
Implementation of BuiltInFunctionDefinitions.SOURCE_WATERMARK.
Channel manager to manage the life cycle of spill channels.
Implementation of BuiltInFunctionDefinitions.SPLIT.
Driver to load the ScriptRunner and execute the script.
Built-in scalar runtime functions.
Utilities for JSON functions.
 
Path spec has two different modes: lax mode and strict mode.
String Like util: Matches a chained sequence of checkers.
Implementation of BuiltInFunctionDefinitions.STARTS_WITH.
Utility to create a StateTtlConfig object.
A DataView which is implemented using state backend.
This interface contains methods for registering StateDataView with a managed store.
Context to switch current key in state backend.
StateListView is a ListView which is implemented using state backends.
StateListView.KeyedStateListView is an default implementation of StateListView whose underlying representation is a keyed state.
StateListView.NamespacedStateListView is an StateListView whose underlying representation is a keyed and namespaced state.
MapView which is implemented using state backends.
A state MapView which does not support nullable keys and namespace.
A state MapView which supports nullable keys but does not support namespace.
A state MapView which supports namespace but does not support nullable keys.
A state MapView which supports nullable keys and namespace.
Streaming unbounded Join operator which supports INNER/LEFT/RIGHT/FULL JOIN.
Streaming unbounded Join operator which supports SEMI/ANTI JOIN.
Wrapper around an Output for wrap T to StreamRecord.
Operator which sets the timestamp on the StreamRecord from the corresponding column of each row.
Operator for stream sort.
Converter for CharType/VarCharType of byte[] external type.
Serializer for StringData.
Serializer configuration snapshot for compatibility and format evolution.
TypeInformation for StringData.
Converter for CharType/VarCharType of String external type.
String utf-8 utils.
Converter for StructuredType of its implementation class.
A processor that processes elements for windows.
Information available in an invocation of methods of SyncStateWindowProcessor.
A rowtime window rank processor.
A WindowOperator for grouped and windowed table aggregates.
The base class for handling table aggregate functions.
The basic implementation of collector for TableFunction.
The basic implementation of collector for ResultFuture in table joining.
TableOperatorWrapper<OP extends org.apache.flink.streaming.api.operators.StreamOperator<org.apache.flink.table.data.RowData>>
This class handles the close, endInput and other related logic of a StreamOperator.
The edge connecting two TableOperatorWrappers.
A generator that generates a TableOperatorWrapper graph from a graph of Transformation.
Table operator to invoke close always.
The operator to temporal join a stream on processing time.
The operator for temporal join (FOR SYSTEM_TIME AS OF o.rowtime) on row time, it has no limitation about message types of the left input and right input, this means the operator deals changelog well.
Converter for TimeType of LocalTime external type.
Converter for TimeType of Long external type.
Serializer for TimestampData.
TypeSerializerSnapshot for TimestampDataSerializer.
TypeInformation for TimestampData.
Converter for TimestampType of LocalDateTime external type.
Converter for TimestampType of Timestamp external type.
Converter for TimeType of Time external type.
A Window that represents a time interval from start (inclusive) to end (exclusive).
The serializer used to write the TimeWindow type.
Serializer configuration snapshot for compatibility and format evolution.
Time util to deals window start and end in different timezone.
TopNBuffer stores mapping from sort key to records list, sortKey is RowData type, each record is RowData type.
A common cache removal listener for rank node.
An implementation of RecordsCombiner that save topN records of incremental input records into the window state.
Factory to create TopNRecordsCombiner.
Implementation of BuiltInFunctionDefinitions.TO_TIMESTAMP_LTZ.
Implementation of BuiltInFunctionDefinitions.TRANSLATE.
Trigger<W extends Window>
A Trigger determines when a pane of a window should be evaluated to emit the results for that part of the window.
A context object that is given to Trigger methods to allow them to register timer callbacks and deal with state.
A GroupWindowAssigner that windows elements into fixed-size windows based on the timestamp of the elements.
Utils for type.
Deprecated.
Use InternalTypeInfo.of(LogicalType) instead if TypeInformation is really required.
Deprecated.
Implementation of BuiltInFunctionDefinitions.TYPE_OF.
The operator for unaligned window table function.
The UnboundedFollowing window frame.
The unbounded window frame calculates frames with the following SQL forms: ...
The UnboundedPreceding window frame.
Implementation of BuiltInFunctionDefinitions.UNHEX.
A special operator which collects all inputs' records and forwards them in MultipleInputStreamOperatorBase.
Flattens ARRAY, MAP, and MULTISET using a table function.
Table function that unwraps the elements of a collection (array or multiset).
Table function that unwraps the elements of a map.
A UnsliceAssigner assigns each element into a single window and not divides the window into finite number of non-overlapping slice.
Utilities to create UnsliceAssigners.
The UnsliceAssigner for session windows.
The UnsliceAssigner for elements have been merged into unslicing windows and attached window start and end timestamps.
An window aggregate processor implementation which works for UnsliceAssigner, e.g.
The UnslicingSyncStateWindowProcessor is an optimized processing for unaligned windows.
A WindowTimerService for unslicing window.
An implementation of RowData which is backed by a RowData and an updated Java object array.
A TopN function could handle updating stream.
Implementation of BuiltInFunctionDefinitions.URL_DECODE.
Implementation of BuiltInFunctionDefinitions.URL_ENCODE.
Generated ValuesInputFormat.
changing rank limit depends on input.
A stream operator that extracts timestamps from stream elements and generates periodic watermarks.
The WatermarkGenerator is used to generate watermark based the input elements.
A function wrapper for FunctionContext if the source supports SupportsWatermarkPushDown.
A Window is a grouping of elements into finite buckets.
We divide windows into 2 categories: Aligned Windows and Unaligned Windows.
 
WindowAssigner is used to assign windows to elements.
A base interface for manipulate async state with window namespace.
A base interface for manipulate state with window namespace.
A wrapper of ValueState which is easier to update based on window namespace.
A buffer that buffers data in memory and flushes many values to state together at a time to avoid frequently accessing state, or flushes to output to reduce shuffling data.
A factory that creates a WindowBuffer.
A factory that creates a WindowBuffer.
A binary map in the structure like Map<WindowKey, BinaryRowData>.
A binary map in the structure like Map<WindowKey, List<BinaryRowData>>.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
A helper to do the window join operations for WindowJoinOperator and AsyncStateWindowJoinOperator.
A streaming window join operator implemented by sync state api.
The WindowJoinOperatorBuilder is used to build a WindowJoinOperator for window join.
The WindowKey structure represents a combination of key and window.
Serializer for WindowKey which is composite with a BinaryRowData and a long value.
A TypeSerializerSnapshot for WindowKeySerializer.
A wrapper of ListState which is easier to update based on window namespace.
A wrapper of ListState which is easier to update based on window namespace.
A wrapper of MapState which is easier to update based on window namespace.
An operator that implements the logic for windowing based on a GroupWindowAssigner and Trigger.
The WindowOperatorBuilder is used to build WindowOperator fluently.
The builder which is used to build AggregateWindowOperator fluently.
The builder which is used to build TableAggregateWindowOperator fluently.
A base window processor provides common methods used for SyncStateWindowProcessor and AsyncStateWindowProcessor.
Information available in an invocation of methods of WindowProcessor.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
The WindowRankOperatorBuilder is used to build a WindowAggOperator for window rank.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
Assigning windows from the sorted input buffers.
Deprecated.
The POJOs in this package are used to represent the deprecated Group Window feature.
A base interface for manipulate state with window namespace.
The WindowTableFunctionOperatorBase acts as a table-valued function to assign windows for input row.
Interface for working with window time and timers which considers timezone for window splitting.
A base implement of WindowTimerService.
A WindowTrigger determines when a window should be evaluated to emit the results.
A wrapper of ValueState which is easier to update based on window namespace.
WrappedRowIterator<T extends org.apache.flink.table.data.RowData>
Wrap MutableObjectIterator to java RowIterator.
A Collector that wraps another collector.
xxHash64.
Converter for YearMonthIntervalType of Period external type.