LoginSign Up
← Back to all sparks
Apache Pinot logo

Apache Pinot

INFRA · APISANALYTICS
Velocity6.3

Real-time distributed OLAP datastore

pinot.apache.org ↗

Pinot 1.6 adds SQL DDL and vector backends, and now needs Java 25 to run its services.

real-time-olapsql-ddlvector-searchmaterialized-viewsapache-projectjava-25
◆Current state
1.6.0 on 29 October lets operators manage tables through CREATE, DROP and SHOW statements, gives materialized views DDL and single-stage query rewrite, and adds GROUPING SETS, ROLLUP and CUBE to both query engines. Vector search gains IVF_FLAT, IVF_PQ and IVF_ON_DISK backends alongside HNSW. The release also removes legacy pieces: native text and FST indexes, Spark 2 and Presto integrations, and APIs deprecated through 1.3.0.
◆Where it's heading
Pinot is moving toward being operated like a SQL database. The multi-stage engine's SQL coverage has grown release by release, and table lifecycle now joins it, while vector backends pull retrieval workloads into the same store. The upgrade cost is real: Java 25 for services, Scala 2.13 for Spark artifacts, and mixed-version limits around SMART aggregation. Federation, the headline of 1.5, doesn't feature in the 1.6 highlights.
◆Prediction
Expect materialized-view rewrite to reach the multi-stage engine, since it is single-stage only today. A security patch on 1.6 would match what followed 1.5.0.

◆Recent moves

  1. 12d ago

    Pinot 1.6.0: SQL DDL, IVF vector search, Java 25 requirement

    ⚡ SPARK

    The first minor since 1.5.0's federation work moves table management into SQL and broadens both analytics and vector search. It also clears out legacy integrations, so it is a far heavier upgrade than the 1.5.1 patch.

    View source ↗
  2. 18d ago

    Pinot 1.6.0 rc1 tag, no release notes

    A bare release-candidate tag ahead of the 1.6.0 final, with nothing beyond the Maven copy message.

    View source ↗
  3. 24d ago

    Pinot 1.6.0 RC0: first candidate with the 1.6 highlights

    The first candidate published the 1.6 feature list for testing: SQL DDL, GROUPING SETS, IVF vector backends and new data types. Its feature highlights carried through to the final release twelve days later.

    View source ↗
  4. 4mo ago

    Apache Pinot 1.5.1

    A security-only patch on 1.5.0: dependency updates and exclusions resolving reported CVEs, with no functional, API, configuration, or wire-format changes and a scan reporting no critical or high findings. Nothing new to use, but a clean upgrade path for operators with vulnerability requirements.

    View source ↗
  5. 6mo ago

    Apache Pinot Release 1.5.0

    ⚡ SPARK

    1.5.0 introduces a federation and multi-cluster routing framework alongside multi-stage engine and upsert work. Federation changes the unit of deployment from a cluster to a fleet, which is a structural shift rather than another round of query features.

    View source ↗
  6. 1y ago

    Apache Pinot Release 1.4.0

    1.4.0 advances the multistage engine with a Lite Mode beta, plus pauseless consumption, logical tables, upsert and dedup work, minion jobs, and rebalancing. Broad and incremental — the groundwork the federation framework builds on a release later.

    View source ↗