What pandas-dev/pandas shipped
Written by FoxPlug from public releases; not affiliated with pandas. An automatic summary of the public release, pull request and commit data of github.com/pandas-dev/pandas. pandas did not write it and does not use or endorse FoxPlug. Every line links to the public change it describes.
Get a weekly update like this for your product, free
Week of September 21, 2026
What shipped
- Fixed to_csv with quoting parameter writing float32 values with float64 representation. Pull request #69490
- Added clearer error message when Series is constructed from a DataFrame, suggesting df.iloc[:, 0]. Pull request #69491
- Fixed to_string formatters adding unwanted leading spaces to object and nullable columns. Pull request #69627
- Fixed default datetime/period formats to zero-pad years before 1000 for consistent output. Pull request #69635
- Float-valued index and Series names now respect display.precision setting. Pull request #69640
- GroupBy.min/max on Arrow date, time and binary columns now use PyArrow's group_by for better performance. Pull request #69488
- Fixed select_dtypes failing to recognize int64 results from numexpr due to nonstandard type representation. Pull request #69427
- read_excel now infers engine from workbook objects (openpyxl, pyxlsb, python_calamine, odf) without requiring explicit engine parameter. Pull request #69457
- Stata reader error message now names the value label set key, which differs from column names in Stata files. Pull request #69507
- Fixed RangeIndex.symmetric_difference incorrectly skipping sort when one side is empty and the other is descending. Pull request #68358
Why it matters
This week brings formatting consistency improvements for numeric types, clearer error messages for common mistakes, and better engine inference for Excel files. Performance optimizations for grouped operations on Arrow data and fixes for type recognition issues address long-standing bugs that affected data integrity and usability.
Changelog entry
- Bug fixes: to_csv with quoting now preserves float32 representation; to_string formatters no longer add leading spaces to object columns Pull request #69490
- Enhancements: Series(DataFrame) raises clearer error; read_excel infers engine from workbook objects Pull request #69491
- Bug fixes: datetime/period default formats zero-pad years before 1000; float index/Series names respect display.precision Pull request #69635
- Performance: GroupBy.min/max on Arrow date, time, binary columns use PyArrow's group_by Pull request #69488
- Bug fixes: select_dtypes now recognizes int64 from numexpr; RangeIndex.symmetric_difference handles empty sides correctly Pull request #69427
- Documentation: Stata reader error messages clarified to show value label set names; plotting backend documentation updated Pull request #69507
This week: clearer errors for Series(DataFrame), consistent float formatting in CSV, better groupby performance on Arrow data, and automatic engine detection for Excel workbooks.
This week's pandas updates focus on output consistency and user experience. Float values now format correctly in CSV with quoting, index names respect display settings, datetime years zero-pad properly, and formatters work uniformly across column types. Performance improvements route Arrow date/time/binary groupby operations through native PyArrow. Engine detection for read_excel now works with openpyxl, pyxlsb, and python_calamine workbook objects.