org.apache.flink.table.functions.hive

接口 HiveFunction

所有已知实现类:

HiveGenericUDAF, HiveGenericUDF, HiveGenericUDTF, HiveScalarFunction, HiveSimpleUDF

@Internal
public interface HiveFunction

Interface for Hive simple udf, generic udf, and generic udtf. TODO: Note: this is only a temporary interface for workaround when Flink type system and udf system rework is not finished. Should adapt to Flink type system and Flink UDF framework later on.

方法概要

限定符和类型	方法和说明
org.apache.flink.table.types.DataType	getHiveResultType(Object[] constantArguments, org.apache.flink.table.types.DataType[] argTypes) Get result type by arguments and argTypes.
void	setArgumentTypesAndConstants(Object[] constantArguments, org.apache.flink.table.types.DataType[] argTypes) Set arguments and argTypes for Function instance.

方法详细资料

setArgumentTypesAndConstants

void setArgumentTypesAndConstants(Object[] constantArguments,
                                  org.apache.flink.table.types.DataType[] argTypes)

Set arguments and argTypes for Function instance. In this way, the correct method can be really deduced by the function instance.

参数:

constantArguments - arguments of a function call (only literal arguments are passed, nulls for non-literal ones)

argTypes - types of arguments

getHiveResultType

org.apache.flink.table.types.DataType getHiveResultType(Object[] constantArguments,
                                                        org.apache.flink.table.types.DataType[] argTypes)

Get result type by arguments and argTypes.

We can't use getResultType(Object[], Class[]). The Class[] is the classes of what is defined in eval(), for example, if eval() is "public Integer eval(Double)", the argTypes would be Class[Double]. However, in our wrapper, the signature of eval() is "public Object eval(Object... args)", which means we cannot get any info from the interface.

参数:

constantArguments - arguments of a function call (only literal arguments are passed, nulls for non-literal ones)

argTypes - types of arguments

返回:

result type.