DuckDB is a high-performance analytical database system. It is designed to be fast, reliable and easy to use. DuckDB provides a rich SQL dialect that supports far more than basic SQL. DuckDB supports arbitrary and nested correlated subqueries, complex types, and more.

The new release contains many improvements to the storage system, general performance improvements, memory management improvements, and new features. Below are the most impactful changes, along with linked PRs that implement them.

Feature change

• Parallel unordered preservation of CREATE TABLE AS and INSERT INTO #5033
• Parallel order save for CREATE TABLE AS and INSERT INTO #5082
• Introduce FSST compression #4366
• Introduce CHIMP128 compression #4878
• Introduce Patas compression (float/double) (Chimp variant) #5044
• Parallel CSV reader #5194
• Parallelize CREATE INDEX of ART #4655
• Improve memory management of ART index #5292
• Nested “UNION” types #4966
• Allow queries to start with FROM instead of SELECT #5076
• Support COLUMNS expression, allowing extended calculation of multiple columns #5120
• Python-style list comprehension syntax #4926
• Out-of-Core Hash improvements #4970
• The jemalloc “extension” for Linux #4971
• Improve rendering of shell result sets #5140
• Add autocompletion support to the shell #4921
• Better looking progress bar #5187

More content can be found in the Release Note and the official blog.

The name “Oxyura” for this version is derived from the bald duck (Oxyura leucocephala), an endangered species native to Eurasia.