Release of SQLite DBMS 3.32. The DuckDB project develops a version of SQLite for analytical queries.

Published release SQLite 3.32.0, a lightweight DBMS designed as a plug-in library. The SQLite code is released into the public domain, meaning it can be used without restrictions and free of charge for any purpose. Financial support for SQLite developers comes from a consortium that includes companies such as Adobe, Oracle, Mozilla, Bentley, and Bloomberg.

Key changes:

• Implemented approximated a variant of the ANALYZE command that allows very large databases to work with partial statistics collection without a full index scan. The limit on the number of records when scanning a single index is set using the new directive "PRAGMA analysis_limit«.
• A new virtual table "bytecode" has been added, providing information about bytecode of prepared statements (prepared statement).
• A VFS layer has been added checksum, adding 8-byte checksums to the end of each data page in the database and checking them upon every read from the database. The layer allows detecting database corruption due to random bit distortions in storage devices.
• A new SQL function iif(X,Y,Z), returning the value Y if expression X is true, or Z otherwise.
• In INSERT and UPDATE statements, column type affinity ( is conducted column affinity) is now enforced before evaluating conditions in the CHECK block.The parameter limit has been increased from 999 to 32766. An extension has been added.
• UINT collating sequence
• with an implementation of sorting sequences that accounts for integers in the text to sort that text in numerical order. In the command line interface, options "—csv", "—ascii", and "—skip" have been added to the ".import" command. Multiple LIKE patterns are allowed in the ".dump" command, merging all tables that match the specified masks in the output. A ".oom" command has been added for debug builds. The ".excel", ".output", and ".once" commands now include the "—bom" option. The ".filectrl" command has been supplemented with the "—schema" option. The expression ESCAPE specified with the LIKE operator now overrides wildcard characters, aligning with PostgreSQL behavior.
• Additionally, the development of the new DBMS
• DuckDB

, which is developing an SQLite variant optimized for executing analytical queriesBesides the shell code from SQLite, the project uses the parser extracted into a separate library from PostgreSQL, the Date Math component from MonetDB.
, its own implementation of window functions (based on the Segment Tree Aggregation algorithm), a vectorized query execution engine (based on the Hyper-Pipelining Query Execution algorithm), and a regular expression handler based on the RE2, its own implementation of window functions (based on the Segment Tree Aggregation algorithm), a vectorized query execution engine (based on the Hyper-Pipelining Query Execution algorithm), a regular expression handler based on the library RE2, its own query optimizer and a Multi-Version Concurrency Control (MVCC) mechanism for managing concurrent task execution.
Project Code is distributed under the MIT license. Development is still at the formation of experimental releases.

Source: opennet.ru