StarRocks' Primary Key table is specifically engineered to deliver powerful query performance during continuous high-volume real-time data ingestion. The primary key index enables data change operations to be resolved at data ingestion, leaving only the latest version of records.

Our massively parallel processing (MPP) architecture with in-memory data shuffling prevents a single node from bottlenecking the entire system, allowing for near-linear scale, especially for JOINs and complex aggregation queries.

In addition to being fully SIMD optimized, StarRocks' own columnar storage engine and vectorized query executions deliver the industry's fastest query performance.

StarRocks' cost-based optimizer is custom-made and deeply optimized with its storage engine. It uses statistics collected from the data to estimate the costs of different execution plans for a query and picks the most efficient query plan.

For extremely high concurrency scenarios, instead of performing denormalization in a separate process, StarRocks' primary key table allows you to update part of your column with the primary key index.

The synchronized materialized view (MV) of StarRocks can be incrementally updated at data ingestion and perform query rewrites at query execution. With guaranteed consistency between the base table and the MV, StarRocks' MV further accelerates real-time analytics.