APACHE-HUDI

Streaming upserts data lake format from Uber.

Definition

Hudi supports two table types: Copy-on-Write (COW, read-optimised) and Merge-on-Read (MOR, write-optimised). Includes incremental queries, indexing (Bloom, HBase, simple), record-level indexes. Compatible with Spark, Flink, Presto, Hive.

Origin

Created by Uber in 2016 ; Apache top-level project June 2020 ; v1.0 released June 2024.

Example in context

Robinhood uses Hudi to ingest trade events in real-time with transparent SQL upserts for analytics.

Related terms

• Delta Lake — competitor.