Package org.apache.hudi.hadoop.realtime

Interface Summary

Interface	Description
RealtimeSplit	Realtime Input Split Interface.

Class Summary

Class	Description
AbstractRealtimeRecordReader	Record Reader implementation to merge fresh avro data with base parquet data, to support real time queries.
HoodieCombineRealtimeRecordReader	Allows to read multiple realtime file splits grouped together by CombineInputFormat.
HoodieEmptyRecordReader	Dummy record for log only realtime split.
HoodieHFileRealtimeInputFormat	HoodieRealtimeInputFormat for HUDI datasets which store data in HFile base file format.
HoodieMergeOnReadTableInputFormat	Base implementation of the Hive's FileInputFormat allowing for reading of Hudi's Merge-on-Read (COW) tables in various configurations: Snapshot mode: reading table's state as of particular timestamp (or instant, in Hudi's terms) Incremental mode: reading table's state as of particular timestamp (or instant, in Hudi's terms) External mode: reading non-Hudi partitions
HoodieParquetRealtimeInputFormat	Input Format, that provides a real-time view of data in a Hoodie table.
HoodieParquetSerde	Simply extends ParquetHiveSerDe.
HoodieRealtimeBootstrapBaseFileSplit	Realtime FileSplit with external base file NOTE: If you're adding fields here you need to make sure that you appropriately de-/serialize them in RealtimeSplit.readFromInput(DataInput) and RealtimeSplit.writeToOutput(DataOutput)
HoodieRealtimeFileSplit	FileSplit implementation that holds Split corresponding to the base file List of HoodieLogFile that holds the delta to be merged (upon reading) This split is correspondent to a single file-slice in the Hudi terminology.
HoodieRealtimePath	Path implementation encoding additional information necessary to appropriately read base files of the MOR tables, such as list of delta log files (holding updated records) associated w/ the base file, etc.
HoodieRealtimeRecordReader	Realtime Record Reader which can do compacted (merge-on-read) record reading or unmerged reading (parquet and log files read in parallel) based on job configuration.
HoodieVirtualKeyInfo	Class to hold virtual key info when meta fields are disabled.