Bodo 2025.11 Release (Date: 11/05/2025)

✨ New Features

• BodoSeries.ai.tokenize now accepts a tokenizer object in addition to a tokenizer factory function.
• Added support for timestamps second, microsecond, and millisecond resolution in BodoDataFrame expressions such as filters.

⚙️ Dependency Changes

• Upgraded PyArrow dependency to 22.0.

2025.11.1

✨ New Features

• Added support for creating empty DataFrames.
• Added support for drop_duplicates() for non-trivial Indexes.
• Added support for any ddof arguments to Series.std().
• Added support for timezone aware types.
• Added support for groupby.size() without value columns.

🏎️ Performance Improvements

• Allow column pruning to function in combination with common table expressions (CTEs) which improves some TPCH queries by 4x.

🐛 Bug Fixes

• Fixed hang related to aggregate operations when amounts of data are similar to the number of workers.
• Fixed non-typical usages of DataFrame.loc objects.
• Enabled accessing multiple Iceberg glue catalogs in the same program using the glue.id parameter with pd.read_sql_table JIT API.
• Fixed incorrect drop_duplicates output when keep=none.
• Fixed error in pd.Timestamp.now and other static timestamp methods.

⚙️ Dependency Changes

• Upgraded the planner for the DataFrame library to the newest release.