Hadi Abdi Khojasteh - Pandas Roadmap and Beyond [PyData Prague #19]

Pandas is migrating core functionality to use PyArrow for improved performance and memory efficiency, especially for string operations and data loading

Copy-on-write functionality has been added in Pandas 2.2, requiring explicit opt-in, which helps clarify data mutations and reduce memory usage

New datetime handling improvements allow better support for different timestamp resolutions beyond just nanoseconds

String operations are being modernized to use PyArrow string types by default instead of NumPy object arrays, providing better performance

New Pandas Enhancement Proposals (PEDEPs) system established to govern future development and community contributions

CSV and JSON readers now dispatch to PyArrow readers, providing up to 10x faster performance

Memory footprint reductions of up to 50% possible with PyArrow implementations

Warning system for chain operations being updated - will show warnings first, then errors in Pandas 3

Community of Pandas contributors growing significantly, with 162 contributors to recent releases

In-place operations being removed or limited to improve consistency and reduce confusion around data mutations