Package com.databricks.jdbc.api.impl.streaming

Class InlineArrowResponseProcessor

java.lang.Object

com.databricks.jdbc.api.impl.streaming.InlineArrowResponseProcessor

All Implemented Interfaces:

ThriftResponseProcessor<ArrowResultChunk>

public class InlineArrowResponseProcessor
extends Object
implements ThriftResponseProcessor<ArrowResultChunk>

Processes Thrift responses into Arrow chunks.

This processor converts TFetchResultsResp into ArrowResultChunk for inline Arrow result handling. It caches the Arrow schema from the first response for use in subsequent batches.

Constructor Summary

Constructors

Constructor	Description
InlineArrowResponseProcessor(StatementId statementId)	Creates a new inline Arrow response processor.

Method Summary

Modifier and Type	Method	Description
Consumer<ArrowResultChunk>	getReleaseAction()	Creates a release action for the data type.
StreamingBatch<ArrowResultChunk>	processInitialResponse(TFetchResultsResp response)	Processes the initial response and returns batch data.
StreamingBatch<ArrowResultChunk>	processResponse(TFetchResultsResp response, long batchIndex, long rowOffset)	Processes a fetched response and returns batch data.

Methods inherited from class java.lang.Object

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

Constructor Detail

InlineArrowResponseProcessor

public InlineArrowResponseProcessor(StatementId statementId)

Creates a new inline Arrow response processor.

Parameters:

statementId - The statement ID for logging and chunk creation

Method Detail

processInitialResponse

public StreamingBatch<ArrowResultChunk> processInitialResponse(TFetchResultsResp response)
                                                        throws DatabricksSQLException

Description copied from interface: ThriftResponseProcessor

Processes the initial response and returns batch data.

This method is called once when the streaming provider is created. It may perform additional initialization such as caching schema information.

Specified by:

processInitialResponse in interface ThriftResponseProcessor<ArrowResultChunk>

Parameters:

response - The initial Thrift fetch response

Returns:

A StreamingBatch containing the processed data

Throws:

DatabricksSQLException - if processing fails

processResponse

public StreamingBatch<ArrowResultChunk> processResponse(TFetchResultsResp response,
                                                        long batchIndex,
                                                        long rowOffset)
                                                 throws DatabricksSQLException

Description copied from interface: ThriftResponseProcessor

Processes a fetched response and returns batch data.

This method is called for each subsequent batch fetched from the server.

Specified by:

processResponse in interface ThriftResponseProcessor<ArrowResultChunk>

Parameters:

response - The Thrift fetch response

batchIndex - The zero-based batch index

rowOffset - The starting row offset for this batch

Returns:

A StreamingBatch containing the processed data

Throws:

DatabricksSQLException - if processing fails

getReleaseAction

public Consumer<ArrowResultChunk> getReleaseAction()

Description copied from interface: ThriftResponseProcessor

Creates a release action for the data type.

This action is called when the batch is released from memory. For types that require explicit cleanup (like ArrowResultChunk with native memory), this should perform that cleanup. For types that don't need cleanup (like ColumnarRowView), this can be a no-op.

Specified by:

getReleaseAction in interface ThriftResponseProcessor<ArrowResultChunk>

Returns:

A Consumer that releases the data