RFC-0066: Native execution support for Hudi MOR tables

[jja725]

Summary

This RFC proposes native execution (Prestissimo/Velox) support for snapshot SELECT queries on Hudi MERGE_ON_READ tables. The main piece is a native Hudi connector in Velox, built around a C++ HoodieLogFormat reader and record merger, structured like the existing Iceberg connector. It also covers the Presto native protocol, explicit Hudi handle conversion, coordinator-side compatibility validation, and native catalog registration.

The revised design treats Java Hudi behavior as the correctness oracle while acknowledging that native semantic parity will be incremental. Each release declares a compatibility matrix for table versions, log formats, merge modes, payloads, key generation, and schema evolution. Unsupported table configurations fail during planning, and unsupported block details discovered on the worker fail the query; there is no JVM fallback from native execution.

Other key points:

• Data predicates remain above the merged scan, matching the existing JVM connector behavior.
• The initial in-memory merge index is visible to Velox memory accounting but non-reclaimable, so native splits have a coordinator- and worker-enforced log-byte limit.
• The pure-C++ implementation includes a maintenance and Hudi-upgrade policy backed by JVM/native differential fixtures.

Related:

• [native]: Hudi MOR (merge-on-read) support presto#28148: Presto implementation tracker
• Add native Hudi connector with MERGE_ON_READ (MOR) snapshot read support facebookincubator/velox#18530: Velox Hudi connector proposal
• [Tracking] Add Avro reader support facebookincubator/velox#18228: Velox Avro reader tracker
• [WIP] prototype: Add Avro reader facebookincubator/velox#15790: original Avro reader prototype
• Support for Hudi on Velox Engine apache/hudi#18308: broader Hudi-on-Velox effort