2025)¶

🎉 Highlights¶

This release, we are excited to significantly improve the import time of the Bodo package, as well as introduce new features like Series.where support and lazy BodoScalars.

✨ New Features¶

Bodo DataFrames now imports the JIT compiler lazily only when necessary, which reduces import time substantially.
Support for Series.where().
Series reductions such as “sum” or “max” now produce a BodoScalar that is evaluated lazily and can be used in some operations such as Series.where() and filter expressions without execution.
Optimized support for “not in series” cases like df[~df.A.isin(df.B)] using anti-join.
Support for bodo.pandas uses inside JIT functions.
Anthropic models used through AWS Bedrock now use Anthropic’s messages API to support newer versions of Claude.

🐛 Bug Fixes¶

Fix for join non-equi condition keys that are not part of the output.
Fix for Series expression with non-range Indexes.

🏎️ Performance Improvements¶

Improved the initialization time for cfuncs used in the acceleration of user defined functions in Series.map and DataFrame.apply calls.

⚙️ Dependency Changes¶

Added upper bound to Numba dependency to avoid issues with version 0.62.