DuckDB version 1.2.0 introduces significant enhancements focused on improving interactions with Apache Arrow. Key features include the ability to push dynamically generated join filters through SQL functions such as UNION, UNNEST, and AGGREGATE, along with bug fixes for arrow table filters and improved scan capabilities for objects interfacing with Python. Notably, the version supports Arrow Non-Canonical Extensions and integrates Arrow query results for optimized data fetching. With these updates, DuckDB aims to enhance performance and interoperability with machine learning workflows and big data systems, particularly for Python users who leverage Arrow's capabilities.