Apache Hudi alternative — Apache Hudi vs Skippr
Apache Hudi is an Apache-2.0 lake table format built around record-key upserts. Current tip is 1.2.1 (25 Sep 2026), a bug-fix on 1.2.0 (table version 9). Copy-on-Write rewrites Parquet file groups on update. Merge-on-Read writes deltas to log files and compacts later. Record-level index (and bucket indexes) steer updates. 1.2.0 added VECTOR, VARIANT, and BLOB types, Spark 4.1 and Flink 2.1 bundles, and Flink record-level index so the index lives in the table, not Flink state. Writers are Spark, Flink (async compaction in the same job), Hudi Streamer (continuous Kafka/DFS/CDC sources), and a Kafka Connect sink. Trino connector for 1.2.1 is in-repo against Trino 483. CDC is not Hudi. A Flink CDC job, Debezium, or Streamer source still captures the WAL. There is no licence fee. You pay Spark/Flink, object storage, timeline services (compaction, clustering, cleaning), and people. This is not Iceberg ∪ WAL. Serving state is Hudi file groups plus logs.
Skippr is an open Streamhouse you run as one Rust binary. Write skippr.yml (or the same thing in Python), discover, sync. You query live state over Iceberg ∪ WAL, and the history is ordinary Iceberg in your bucket. No Hudi timeline. ELT into a warehouse is one path, not the product.
| Feature | Skippr | Apache Hudi |
|---|---|---|
| Deploy | Single Rust binary, your VPC or on-prem | Spark, Flink, Streamer, or Connect on a lake |
| Schema | In-stream evolution, sibling columns | Hudi schema evolution. VARIANT/VECTOR/BLOB on 1.2 |
| Storage | Open Iceberg on Parquet you own | Hudi CoW or MoR on your bucket. Not Iceberg |
| CDC | Log-based; fails closed if the dest cannot honour | Streamer / Flink CDC / Debezium into Hudi |
| Query | skipprd query on Iceberg ∪ WAL; history in any Iceberg engine | Spark/Flink/Trino with a Hudi connector. Incremental by commit time |
| Pricing | Free self-hosted engine | Apache-2.0. Pay the writer cluster and table services. |
What Apache Hudi is good at
Continuous upserts into a lake by record key, with incremental pull by commit time, compaction inside the writer instead of a separate Spark job. That is the product. Flink MoR plus record-level index is why some streaming estates still pick Hudi over Iceberg appends. If the job is "this changelog into this Hudi table," that is a real product.
Where the job is wider
Hudi is a table format plus table services. It is not a Streamhouse. There is no Iceberg ∪ WAL. There is no skipprd query on in-flight Iceberg rows. Postgres is not a Hudi source. CDC starts in Flink CDC, Debezium, or Streamer. Iceberg engines need a Hudi connector, not the Iceberg REST catalog you already run.
Timeline math is the ops. Compaction, clustering, cleaning, and file-group layout are part of correctness. Skippr will not write Hudi. Skippr also will not replace a Hudi MoR table that already serves incremental Spark jobs. Iceberg Connect is a separate page.
If the job is "Hudi upserts we already staff," Hudi is the narrower tool. If the job is live SQL on Iceberg you run, the format was never that.
Getting started
Install skipprd from the install guide, or pip install skippr. The three tabs write the same skippr.yml. Then discover, sync, and query.
skippr:
workspace: analytics
skipprd_el_storage_mode: local
pipelines:
orders:
data_source: data_sources.app_db
data_sink: data_sinks.lake
data_sources:
app_db:
Postgres:
host: db.internal
user: skippr_reader
password: ${APP_DB_PASSWORD}
database: app
tables: [orders, customers]
cdc_mode: snapshot_then_cdc
data_sinks:
lake:
AthenaIceberg:
warehouse: s3://my-iceberg-warehouse/
glue_database_name: analytics
athena_workgroup_name: primary
athena_results_s3_bucket: my-athena-results
region: us-east-1
schema_sink: schema_sinks.lake_schema
schema_sinks:
lake_schema:
AthenaIceberg:
warehouse: s3://my-iceberg-warehouse/
glue_database_name: analytics
athena_workgroup_name: primary
athena_results_s3_bucket: my-athena-results
region: us-east-1skipprd --workspace analytics --storage-mode local connect data-source postgres \
--pipeline orders \
--name app_db \
--host db.internal \
--user skippr_reader \
--password '${APP_DB_PASSWORD}' \
--database app \
--tables orders \
--tables customers \
--cdc-mode snapshot_then_cdc
skipprd connect data-sink athenaiceberg \
--pipeline orders \
--name lake \
--warehouse s3://my-iceberg-warehouse/ \
--glue-database-name analytics \
--athena-workgroup-name primary \
--athena-results-s3-bucket my-athena-results \
--region us-east-1
skipprd connect schema-sink athenaiceberg \
--pipeline orders \
--name lake_schemafrom skippr import (
Config,
DataSinkAthenaIceberg,
DataSourcePostgres,
DataSourcePostgresSourceCdcMode,
EnvRef,
LocalStorage,
Pipeline,
)
cfg = Config().workspace("analytics").storage(LocalStorage())
app_db = cfg.data_source(
"app_db",
DataSourcePostgres(
host="db.internal",
user="skippr_reader",
password=EnvRef("APP_DB_PASSWORD"),
database="app",
tables=["orders", "customers"],
cdc_mode=DataSourcePostgresSourceCdcMode.SNAPSHOTTHENCDC,
),
)
lake = cfg.data_sink(
"lake",
DataSinkAthenaIceberg(
warehouse="s3://my-iceberg-warehouse/",
glue_database_name="analytics",
athena_workgroup_name="primary",
athena_results_s3_bucket="my-athena-results",
region="us-east-1",
),
schema_sink="lake_schema",
)
cfg.pipeline("orders", Pipeline(data_source=app_db, data_sink=lake))
cfg.save("skippr.yml")export APP_DB_PASSWORD='your-postgres-password'
skipprd discover --pipeline orders
skipprd sync --pipeline orders --once
skipprd query "SELECT count(*) FROM analytics.orders"Which to choose
Choose Hudi if you want record-key upserts in a lake, and Spark or Flink already writes that table.
Choose Skippr if you want one binary, live SQL on Iceberg ∪ WAL, and files that stay yours. Skippr will not run a Hudi writer.
Install from Install. Connector and query reference live in Developer Docs. For optional onboarding in your VPC, Enterprise Jumpstart or Contact.