org.apache.flink.table.runtime.functions.aggregate

Class BuiltInAggregateFunction<T,ACC>

java.lang.Object

org.apache.flink.table.functions.UserDefinedFunction

org.apache.flink.table.functions.ImperativeAggregateFunction<T,ACC>

org.apache.flink.table.functions.AggregateFunction<T,ACC>

org.apache.flink.table.runtime.functions.aggregate.BuiltInAggregateFunction<T,ACC>

All Implemented Interfaces:

Serializable, FunctionDefinition

Direct Known Subclasses:

BatchApproxCountDistinctAggFunctions.ApproxCountDistinctAggFunction, CollectAggFunction, FirstValueAggFunction, FirstValueWithRetractAggFunction, JsonArrayAggFunction, JsonObjectAggFunction, LagAggFunction, LastValueAggFunction, LastValueWithRetractAggFunction, ListAggWithRetractAggFunction, ListAggWsWithRetractAggFunction, MaxWithRetractAggFunction, MinWithRetractAggFunction

@Internal
public abstract class BuiltInAggregateFunction<T,ACC>
extends AggregateFunction<T,ACC>

Base class for runtime implementation represented as AggregateFunction that is constructed from SpecializedFunction.specialize(SpecializedContext).

Subclasses must offer a constructor that takes SpecializedFunction.SpecializedContext if they are constructed from a BuiltInFunctionDefinition. Otherwise the BuiltInAggregateFunction() constructor might be more appropriate.

By default, all built-in functions work on internal data structures. However, this can be changed by overriding getArgumentDataTypes(), getAccumulatorDataType(), and getOutputDataType(). Or by overriding getTypeInference(DataTypeFactory) directly.

Since the accumulator type is runtime specific, it must be declared explicitly; otherwise it is derived from the output type.

See Also:

Serialized Form

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	BuiltInAggregateFunction()
protected	BuiltInAggregateFunction(BuiltInFunctionDefinition definition, SpecializedFunction.SpecializedContext context)

Method Summary

Modifier and Type	Method and Description
DataType	getAccumulatorDataType()
List<DataType>	getArgumentDataTypes()
DataType	getOutputDataType()
Set<FunctionRequirement>	getRequirements() Returns the set of requirements this definition demands.
TypeInference	getTypeInference(DataTypeFactory typeFactory) Returns the logic for performing type inference of a call to this function definition.
boolean	isDeterministic() Returns information about the determinism of the function's results.

Methods inherited from class org.apache.flink.table.functions.AggregateFunction

getKind, getValue

Methods inherited from class org.apache.flink.table.functions.ImperativeAggregateFunction

createAccumulator, getAccumulatorType, getResultType

Methods inherited from class org.apache.flink.table.functions.UserDefinedFunction

close, functionIdentifier, open, toString

Methods inherited from class java.lang.Object

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

Constructor Detail

BuiltInAggregateFunction

protected BuiltInAggregateFunction(BuiltInFunctionDefinition definition,
                                   SpecializedFunction.SpecializedContext context)

BuiltInAggregateFunction

protected BuiltInAggregateFunction()

Method Detail

getArgumentDataTypes

public List<DataType> getArgumentDataTypes()

getAccumulatorDataType

public DataType getAccumulatorDataType()

getOutputDataType

public DataType getOutputDataType()

getTypeInference

public TypeInference getTypeInference(DataTypeFactory typeFactory)

Description copied from class: UserDefinedFunction

Returns the logic for performing type inference of a call to this function definition.

The type inference process is responsible for inferring unknown types of input arguments, validating input arguments, and producing result types. The type inference process happens independent of a function body. The output of the type inference is used to search for a corresponding runtime implementation.

Instances of type inference can be created by using TypeInference.newBuilder().

See BuiltInFunctionDefinitions for concrete usage examples.

The type inference for user-defined functions is automatically extracted using reflection. It does this by analyzing implementation methods such as eval() or accumulate() and the generic parameters of a function class if present. If the reflective information is not sufficient, it can be supported and enriched with DataTypeHint and FunctionHint annotations.

Note: Overriding this method is only recommended for advanced users. If a custom type inference is specified, it is the responsibility of the implementer to make sure that the output of the type inference process matches with the implementation method:

The implementation method must comply with each DataType.getConversionClass() returned by the type inference. For example, if DataTypes.TIMESTAMP(3).bridgedTo(java.sql.Timestamp.class) is an expected argument type, the method must accept a call eval(java.sql.Timestamp).

Regular Java calling semantics (including type widening and autoboxing) are applied when calling an implementation method which means that the signature can be eval(java.lang.Object).

The runtime will take care of converting the data to the data format specified by the DataType.getConversionClass() coming from the type inference logic.

Specified by:

getTypeInference in interface FunctionDefinition

Overrides:

getTypeInference in class AggregateFunction<T,ACC>

getRequirements

public Set<FunctionRequirement> getRequirements()

Description copied from interface: FunctionDefinition

Returns the set of requirements this definition demands.

isDeterministic

public boolean isDeterministic()

Description copied from interface: FunctionDefinition

Returns information about the determinism of the function's results.

It returns true if and only if a call to this function is guaranteed to always return the same result given the same parameters. true is assumed by default. If the function is not purely functional like random(), date(), now(), ... this method must return false.

Furthermore, return false if the planner should always execute this function on the cluster side. In other words: the planner should not perform constant expression reduction during planning for constant calls to this function.