4.8.5
📦 datasetsView on GitHub →
🐛 8 fixes🔧 4 symbols
Summary
This release focuses primarily on bug fixes, addressing issues related to JSON decoding, batching, and remote file handling during preparation. Several CI/CD and internal maintenance tasks were also completed.
🐛 Bug Fixes
- Fixed issue where Json() values were not decoded before calling DataFrame.to_json().
- Fixed JSON type decoding before to_list or to_dict is called.
- Fixed batching for table-formatted datasets.
- Fixed iterable map resume state.
- Prevented embedding remote files in download_and_prepare to parquet.
- Fixed mask in embed_storage for remote files.
- Fixed missing original_files in ci json test.
- Fixed null handling in embed storage.