v5.0.0-beta.6

What's Changed

New Features 🎉

• feat: introduce CacheCodec for serializing index cache entries by @wjones127 in #6223
• feat(java): add allowExternalBlobOutsideBases to WriteParams by @beinan in #6330
• feat: add batch_size_bytes to encoding decode stream by @westonpace in #6388
• feat: validate DataFile column_indices at commit time by @westonpace in #6414
• feat(java): add Dataset.getZonemapStats() API by @beinan in #6421
• feat: support IVF partitions multi-split by @BubbleCal in #6423
• feat(python): add blob-aware to_pandas by @BubbleCal in #6424
• feat: add DataFile.create helper for building DataFile metadata by @westonpace in #6427

Bug Fixes 🐛

• fix: return error instead of OOM on corrupt binary page offsets by @westonpace in #6392
• fix: implement RowIdMeta.asdict() and from_dict() for stable row ID serialization by @pengw0048 in #6405
• fix(namespace): remove trailing empty query in table uri by @jackye1995 in #6415
• fix: skip legacy validation for Overwrite operations by @hamersaw in #6418
• fix: implement next page token support in list namespaces and tables by @ayao227 in #6419
• fix: avoid corrupting sub-schema merge_insert pages on v2.2 by @Xuanwo in #6420
• fix(merge_insert): use sentinel column for NULL-safe source row detection by @pratik0316 in #6439

Documentation 📚

• docs: document DataFile column_indices changes in 2.1 format by @westonpace in #6416
• docs: add community sync to docs by @ccmao1130 in #6434

Full Changelog: v5.0.0-beta.5...v5.0.0-beta.6