What's Changed
- Add release notes for 2025.7.4 by @IsaacWarren in #650
- [BSE-4947] Support Series.std and Series.describe by @chrisoh7 in #644
- Avoid vendoring mpi4py in conda env by @scott-routledge2 in #636
- Treat missing index columns in Pandas Metadata as RangeIndex when reading Parquet by @scott-routledge2 in #652
- Fix join output column map to input by @ehsantn in #654
- [BSE-4946] Suppress subsequent fallback warnings by @chrisoh7 in #651
- Refactor plan code into a new file by @ehsantn in #656
- [BSE-4939] Spark Iceberg benchmark by @chrisoh7 in #649
- Todd/tpch3 by @DrTodd13 in #643
- Fix DateTimeProperties validation by @chrisoh7 in #660
- Refactor LazyPlan uses into proper Python class hierarchy by @ehsantn in #659
- [BSE-4978] Support Series.aggregate by @chrisoh7 in #657
- [BSE-4949] Support column selection in DataFrameGroupby getattr by @chrisoh7 in #653
- Handle Parquet read filter columns not used elsewhere in the query by @ehsantn in #664
- Fix df setitem for common Series.str.cat cases by @ehsantn in #665
- join expr null handling by @DrTodd13 in #662
- Update docs by @chrisoh7 in #663
- BSE-4975: Change parquet readahead by @IsaacWarren in #655
- NYC taxi changes by @scott-routledge2 in #658
Full Changelog: 2025.7.4...2025.7.5