After six years of development presented DBMS release DuckDB 1.0, positioned as a variant of SQLite for analytical queries. DuckDB combines the features of SQLite, such as compactness, the ability to connect in the form of an embedded library, storing the database in a single file, and a convenient CLI interface, with tools and optimizations for execution analytical queries, covering a significant portion of the stored data, for example, performing an aggregation of the entire contents of tables or merging several large tables. The project code is written in C++ and distributed by under MIT license.
Version 1.0 is marked as the first stable release of the project, the preparation of which focused on improving stability rather than increasing functionality. The new version also fixes the data storage format, which has been backward compatible since the previous release. In subsequent releases, the developers intend to be more careful about adding new features and take care to maintain compatibility between releases, as well as stabilize the SQL dialect and C API. If a change in SQL semantics is required in the future, developers will publish a warning in advance about upcoming changes and provide workarounds to preserve the functionality of the existing code.
DuckDB supports an extended dialect of the SQL language that includes additional capabilities for processing very complex and time-consuming queries. For example, it is possible to use complex types (arrays, structures, unions), as well as perform arbitrary and nested correlating subqueries. Supports simultaneous execution of multiple queries, executing queries directly from CSV and Parquet. Support for import from PostgreSQL DBMS is available.
In addition to the shell code from SQLite, the project uses a parser from PostgreSQL and a Date Math component from MonetDBits own implementation of window functions (based on the Segment Tree Aggregation algorithm), a regular expression processor based on the RE2 library, its own query optimizer, an MVCC mechanism for managing the simultaneous execution of tasks (Multi-Version Concurrency Control), as well as a vectorized query execution engine based on the algorithm Hyper-Pipelining Query Execution, which allows you to process large sets of values at once in one operation.
Thanks for reading: