Export - CSV (All fields)
Export - CSV (Current fields)
HUDI-1046
Support updates during clustering in CoW modeHUDI-3187
Initiate RFC Lock free concurrency control
Change the index plan to capture everything that is needed to create index
Enhance retries for failed writes w/ write conflicts in a multi writer scenarios
Write RFC for index functionHUDI-5249
Support MetadataColumnStatsIndex for Spark record
fetchAllLogsMergedFileSlice will read basefile which it does not expect
Add new log block header type to store record positions
Extend merge API to support all merging operations
Introduce an engine-agnostic FileGroupReader for snapshot read
Integrate FileGroupReader with NewHoodieParquetFileFormat for Spark MOR Snapshot Query
Support incremental read in engine-agnostic FileGroupReader
Support time-travel read in engine-agnostic FileGroupReader
Implement generation of record_positions for updates and deletes on write path
Implement position-based deletes in FileGroupReader
Implement position-based updates in FileGroupReader
Implement log writing with partial updates on the write path
Implement merging of partial updates in FileGroupReader
Make sure serialization of log blocks is language and architecture independent
Use UTF_8 to encode String to byte array in all places
Port BloomFilter related classes from Hadoop library to remove dependency
Add tests and docs for ported Bloom Filter classes
Move configs for reading a file group to hudi-common moduleHUDI-6946
Data Duplicates with range pruning while using hoodie.bloom.index.use.metadata
Fix CI failure on master
Instantiate HoodieFileGroupRecordBuffer inside new file group reader
Turn on new features by default through configs for 1.0.0-beta1
Solid completion time on timeline
Make sure that non-string primary-keys are supported
Clustering MOR applies base files after log files
ExternalSpillableMap may cause data duplication when flink compaction
General Redesign of Archived Timeline for efficient scan and management
Integration flink metric in flink stream
HoodieFlinkWriteClient supports to allow parallel writing to tables using Locking service
Add metrics for async indexer
Clean up indexing APIs in write client
Clean up usages of "assume.date.partition" config within hudi
End-to-end PoC of functional indexes
Add abstraction to track a function defined on a column
Store index function in table properties
Support index function for Spark SQL built-in functions
Add functionality to create new MT partition for index function
Support file pruning based on new index function in Spark
Support "CREATE INDEX" for index function through Spark SQL
Archived Timeline as LSM Tree - Initial Impl
HoodieTableSource supports parquet predicate push down
Support TimelineBased Checkpoint Metadata for flink
support hudi table type changing in hudi-cli module
Flink lockless multi-writer
Fix get valid checkpoint for current writer