apache/hudi

Health v2 · maintenance only

Adoption not part of the score

6219 stars · 2501 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Apache Hudi is an open data lakehouse platform built on a high-performance open table format that brings database functionality like transactions, upserts, deletes, and indexing to data lakes. It supports incremental processing pipelines with low-latency analytics across engines such as Spark, Flink, Presto, and Trino.

Use cases

• upsert and delete records in a data lake on S3
• build incremental streaming pipelines instead of batch jobs
• ingest database CDC changes into a lakehouse table
• query data lake tables with ACID guarantees from Spark or Trino
• time travel to query historical versions of a table
• deduplicate and merge high-throughput streaming data

When to choose

• you need fast, mutable upserts/deletes on cloud object storage
• you want minute-level incremental processing over a data lake
• you need ACID transactions and concurrency control across multiple writers
• you use Spark or Flink and want an open table format with rich indexing

When to avoid

• you only need append-only immutable data with simple reads
• your stack is non-JVM and you need full write support (hudi-rs is still maturing)
• you want a fully managed warehouse without operating table services like compaction and clustering

Facets

library · maturity stable

etl streaming database data-science big-data databases analytics jvm cloud self-hosted data-lakehouse table-format upserts incremental-processing acid-transactions apache-spark apache-flink cdc time-travel open-table-format data-engineering docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="apache/hudi")