Package com.amazonaws.athena.connector.lambda.handlers

Class UserDefinedFunctionHandler

java.lang.Object

com.amazonaws.athena.connector.lambda.handlers.UserDefinedFunctionHandler

All Implemented Interfaces:

com.amazonaws.services.lambda.runtime.RequestStreamHandler

Direct Known Subclasses:

AthenaUDFHandler, ExampleUserDefinedFuncHandler

public abstract class UserDefinedFunctionHandler
extends Object
implements com.amazonaws.services.lambda.runtime.RequestStreamHandler

Athena UDF users are expected to extend this class to create UDFs.

Constructor Summary

Constructors

Constructor	Description
UserDefinedFunctionHandler(String sourceType)

Method Summary

Modifier and Type	Method	Description
protected void	doHandleRequest(BlockAllocator allocator, com.fasterxml.jackson.databind.ObjectMapper objectMapper, UserDefinedFunctionRequest req, OutputStream outputStream)
void	handleRequest(InputStream inputStream, OutputStream outputStream, com.amazonaws.services.lambda.runtime.Context context)
protected void	onPing(PingRequest request)
protected Block	processRows(BlockAllocator allocator, Method udfMethod, Block inputRecords, org.apache.arrow.vector.types.pojo.Schema outputSchema)	Processes a group by rows.

Methods inherited from class java.lang.Object

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

Constructor Detail

UserDefinedFunctionHandler

public UserDefinedFunctionHandler(String sourceType)

Method Detail

handleRequest

public final void handleRequest(InputStream inputStream,
                                OutputStream outputStream,
                                com.amazonaws.services.lambda.runtime.Context context)

Specified by:

handleRequest in interface com.amazonaws.services.lambda.runtime.RequestStreamHandler

doHandleRequest

protected final void doHandleRequest(BlockAllocator allocator,
                                     com.fasterxml.jackson.databind.ObjectMapper objectMapper,
                                     UserDefinedFunctionRequest req,
                                     OutputStream outputStream)
                              throws Exception

Throws:

Exception

processRows

protected Block processRows(BlockAllocator allocator,
                            Method udfMethod,
                            Block inputRecords,
                            org.apache.arrow.vector.types.pojo.Schema outputSchema)
                     throws Exception

Processes a group by rows. This method takes in a block of data (containing multiple rows), process them and returns multiple rows of the output column in a block.

UDF methods are invoked row-by-row in a for loop. Arrow values are converted to Java Objects and then passed into the UDF java method. This is not very efficient because we might potentially be doing a lot of data copying. Advanced users could choose to override this method and directly deal with Arrow data to achieve better performance.

Parameters:

allocator - arrow memory allocator

udfMethod - the extracted java method matching the User-Defined-Function defined in Athena.

inputRecords - input data in Arrow format

outputSchema - output data schema in Arrow format

Returns:

output data in Arrow format

Throws:

Exception

onPing

protected void onPing(PingRequest request)