Package com.databricks.jdbc.api.impl.arrow

Interface ChunkLinkFetcher

All Known Implementing Classes:

SeaChunkLinkFetcher, ThriftChunkLinkFetcher

public interface ChunkLinkFetcher

Abstraction for fetching chunk links from either SEA or Thrift backend. Implementations handle the protocol-specific details of how links are retrieved.

This interface enables a unified streaming approach for chunk downloads regardless of the underlying client type (SEA or Thrift).

Method Summary

Modifier and Type	Method	Description
void	close()	Closes any resources held by the fetcher.
ChunkLinkFetchResult	fetchLinks(long startChunkIndex, long startRowOffset)	Fetches the next batch of chunk links starting from the given position.
ExternalLink	refetchLink(long chunkIndex, long rowOffset)	Refetches a specific chunk link that may have expired.

Method Detail

fetchLinks

ChunkLinkFetchResult fetchLinks(long startChunkIndex,
                                long startRowOffset)
                         throws DatabricksSQLException

Fetches the next batch of chunk links starting from the given position.

The implementation may return one or more links in a single call. The returned ChunkLinkFetchResult indicates whether more chunks are available.

SEA implementations use startChunkIndex while Thrift implementations use startRowOffset. Each implementation uses the parameter relevant to its protocol and ignores the other.

Parameters:

startChunkIndex - The chunk index to start fetching from (used by SEA)

startRowOffset - The row offset to start fetching from (used by Thrift with FETCH_ABSOLUTE)

Returns:

ChunkLinkFetchResult containing the fetched links and continuation information

Throws:

DatabricksSQLException - if the fetch operation fails

refetchLink

ExternalLink refetchLink(long chunkIndex,
                         long rowOffset)
                  throws DatabricksSQLException

Refetches a specific chunk link that may have expired.

This is used when a previously fetched link has expired before the chunk could be downloaded. Both SEA and Thrift clients support this via the getResultChunks API.

SEA uses chunkIndex while Thrift uses rowOffset to identify the chunk to refetch.

Parameters:

chunkIndex - The specific chunk index to refetch (used by SEA)

rowOffset - The row offset of the chunk to refetch (used by Thrift)

Returns:

The refreshed ExternalLink with a new expiration time

Throws:

DatabricksSQLException - if the refetch operation fails

close

void close()

Closes any resources held by the fetcher.