Calcite

Apache Calcite is a dynamic data management framework that provides a SQL parser/validator, a rich relational algebra, and a cost-based query optimizer that can plan across many heterogeneous data sources. It doesn’t store data or execute queries itself; instead it offers adapters to “federate” sources (JDBC databases, NoSQL, files, streaming) and pushes operations like filters, projections, and joins down to each system where possible. Calcite’s rule-driven optimizer and extensible type system let you define custom operators and schemas, and projects like Apache Druid, Apache Hive, Apache Drill, and Apache Phoenix use it to power their SQL layers. Its Avatica subproject adds a lightweight, database-agnostic remote JDBC/HTTP wire protocol for embedding SQL services.

License: Apache 2.0

Tags: Data, Parsing, Validation, Optimization, Queries

Features: SQL parser, SQL validator, relational algebra, cost-based optimizer, rule-based planner, query rewriting, materialized view rewriting, predicate pushdown, join reordering, federated query planning, adapters for heterogeneous data sources, extensible type system, user-defined functions and operators, schema management and virtual schemas, streaming SQL support, semi-structured data support, trait-based optimization framework, enumerable interpreter, Avatica JDBC/HTTP wire protocol, planner and metadata SPI for custom rules and adapters

Website: https://calcite.apache.org/