Summary

This release introduced some excellent new features, including upsert, tiered storage, pinot-spark-connector, support of having clause, more validations on table config and schema, support of ordinals in GROUP BY and ORDER BY clause, array transform functions, adding push job type of segment metadata only mode, and some new APIs like updating instance tags, new health check endpoint. It also contains many key bug fixes. See details below.

The release was cut from the following commit: e5c9bec and the following cherry-picks:

• ​d033a11​

Notable New Features

• Tiered storage (#5793)

• Upsert feature (#6096, #6113, #6141, #6149, #6167)

• Pre-generate aggregation functions in QueryContext (#5805)

• Adding controller healthcheck endpoint: /health (#5846)

• Add pinot-spark-connector (#5787)

• Support multi-value non-dictionary group by (#5851)

• Support type conversion for all scalar functions (#5849)

• Add additional datetime functionality (#5438)

• Support post-aggregation in ORDER-BY (#5856)

• Support post-aggregation in SELECT (#5867)

• Add RANGE FilterKind to support merging ranges for SQL (#5898)

• Add HAVING support (#58895889)

• Support for exact distinct count for non int data types (#5872)

• Add max qps bucket count (#5922)

• Add Range Indexing support for raw values (#5853)

• Add IdSet and IdSetAggregationFunction (#5926)

• [Deepstore by-pass]Add a Deepstore bypass integration test with minor bug fixes. (#5857)

• Add Hadoop counters for detecting schema mismatch (#5873)

• Add RawThetaSketchAggregationFunction (#5970)

• Instance API to directly updateTags (#5902)

• Add streaming query handler (#5717)

• Add InIdSetTransformFunction (#5973)

• Add ingestion descriptor in the header (#5995)

• Zookeeper put api (#5949)

• Feature/#5390 segment indexing reload status api (#5718)

• Segment processing framework (#5934)

• Support streaming query in QueryExecutor (#6027)

• Add list of allowed tables for emitting table level metrics (#6037)

• Add FilterOptimizer which supports optimizing both PQL and SQL query filter (#6056)

• Adding push job type of segment metadata only mode (#5967)

• Minion taskExecutor for RealtimeToOfflineSegments task (#6050, #6124)

• Adding array transform functions: array_average, array_max, array_min, array_sum (#6084)

• Allow modifying/removing existing star-trees during segment reload (#6100)

• Implement off-heap bloom filter reader (#6118)

• Support for multi-threaded Group By reducer for SQL. (#6044)

• Add OnHeapGuavaBloomFilterReader (#6147)

• Support using ordinals in GROUP BY and ORDER BY clause (#6152)

• Merge common APIs for Dictionary (#6176)

• Add table level lock for segment upload ([#6165])

• Added recursive functions validation check for group by (#6186)

• Add StrictReplicaGroupInstanceSelector (#6208)

• Add IN_SUBQUERY support (#6022)

• Add IN_PARTITIONED_SUBQUERY support (#6043)

• Some UI features (#5810, #5981, #6117, #6215)

Special notes

• Brokers should be upgraded before servers in order to keep backward-compatible:

  • Change group key delimiter from '\t' to '\0' (#5858)

  • Support for exact distinct count for non int data types (#5872)

• Pinot Components have to be deployed in the following order:

  (PinotServiceManager -> Bootstrap services in role ServiceRole.CONTROLLER -> All remaining bootstrap services in parallel)

  • Starts Broker and Server in parallel when using ServiceManager (#5917)

  • New settings introduced and old ones deprecated:

  • Make realtime threshold property names less ambiguous (#5953)

  • Change Signature of Broker API in Controller (#6119)

• This aggregation function is still in beta version. This PR involves change on the format of data sent from server to broker, so it works only when both broker and server are upgraded to the new version:

  • Enhance DistinctCountThetaSketchAggregationFunction (#6004)

Major Bug fixes

• Improve performance of DistinctCountThetaSketch by eliminating empty sketches and unions. (#5798)

• Enhance VarByteChunkSVForwardIndexReader to directly read from data buffer for uncompressed data (#5816)

• Fixing backward-compatible issue of schema fetch call (#5885)

• Fix race condition in MetricsHelper (#5887)

• Fixing the race condition that segment finished before ControllerLeaderLocator created. (#5864)

• Fix CSV and JSON converter on BYTES column (#5931)

• Fixing the issue that transform UDFs are parsed as function name 'OTHER', not the real function names (#5940)

• Incorporating embedded exception while trying to fetch stream offset (#5956)

• Use query timeout for planning phase (#5990)

• Add null check while fetching the schema (#5994)

• Validate timeColumnName when adding/updating schema/tableConfig (#5966)

• Handle the partitioning mismatch between table config and stream (#6031)

• Fix built-in virtual columns for immutable segment (#6042)

• Refresh the routing when realtime segment is committed (#6078)

• Add support for Decimal with Precision Sum aggregation (#6053)

• Fixing the calls to Helix to throw exception if zk connection is broken (#6069)

• Allow modifying/removing existing star-trees during segment reload (#6100)

• Add max length support in schema builder (#6112)

• Enhance star-tree to skip matching-all predicate on non-star-tree dimension (#6109)

Backward Incompatible Changes

• Make realtime threshold property names less ambiguous (#5953)

• Enhance DistinctCountThetaSketchAggregationFunction (#6004)

• Deep Extraction Support for ORC, Thrift, and ProtoBuf Records (#6046)