apache/doris

Apache Doris is an open-source, MPP-based database aimed at teams that need fast SQL analytics over freshly arriving data, plus search across structured, text, JSON, and vector columns in the same engine. It is a database to operate, not a library you add to an application; the README positions it around real-time analytics, lakehouse query acceleration, and hybrid search Apache Doris repository.

What changes in a workflow

Instead of separate stacks for dashboards, log analysis, and vector retrieval, Doris presents one SQL surface. Documented use cases include customer-facing sub-second analytics, consolidated real-time warehousing, high-throughput observability data, and AI/search workloads combining vector, text, JSON, and structured filters use cases. The practical consequence is fewer systems to keep consistent, but the engine's capabilities and tuning become the thing you learn.

Ingestion is continuous: the project describes streaming ingestion, incremental transformation, and sub-second queries under high concurrency as one capability real-time analytics. Lakehouse support queries open table formats such as Iceberg, Delta Lake, and Hudi directly, so teams can analyze data already in object storage without first copying everything into Doris-managed storage.

Two deployment shapes

Doris supports compute-storage coupled and decoupled modes. In decoupled mode, stateless compute groups run over shared object storage, which lets you scale compute on demand and isolate workloads deployment mode guide. Decoupled mode introduces dependence on that object storage layer; coupled mode keeps storage with compute. The project points to a dedicated guide for choosing between them rather than claiming one fits all.

Prerequisites and limits

The README is a gateway, not a complete operations manual. Installation, quick start, and Kubernetes operator deployment have separate documentation installation, Kubernetes operator. Source compilation is documented through a Docker-based path rather than presented as a lightweight local build.

Licensing needs attention. Doris itself is Apache-2.0, but the README warns that some third-party dependency licenses are incompatible with Apache 2.0 and that certain Doris features must be disabled to comply, with details in thirdparty/LICENSE.txt license note. That is a packaging/distribution condition to check, not a restriction on ordinary use of the database.

Performance claims here are qualitative ("sub-second", "high concurrency"); the supplied material contains no benchmarks, hardware details, or scale figures. Production adoption is asserted broadly by the project and echoed in a 2022 Hacker News comment claiming 500+ companies, which is dated and self-reported, not independent validation 2022 discussion.

The README shows stars and topics but no version-specific compatibility matrix; check current release notes and deployment docs for supported versions before committing.