v1.0.0-beta.4

What's Changed

Breaking Changes 🛠

• perf!: dynamic pruning for vector search by @BubbleCal in #4773
• feat!: remove unnecessary mut of dataset::sql by @ddupg in #5207
• refactor!: move all previous code into previous mod by @Xuanwo in #5217
• refactor!: deprecate TFRecord support by @jackye1995 in #4593

Critical Fixes ‼️

• fix: convert some panics into errors by @wjones127 in #5258

New Features 🎉

• feat: provide inline_transaction model for IO optimizing by @majin1102 in #4774
• feat(python): add support for HuggingFace IterableDataset by @changhiskhan in #2599
• feat: add public accessors for count plan construction by @wkalt in #5103
• feat: add adapter for REST namespace with manifest namespace backend by @jackye1995 in #4984
• feat: add blob compaction support by @Xuanwo in #5189
• feat: add inline optimization for dir namespace by @jackye1995 in #5244
• feat: conflict resolution for DataReplacement by @wjones127 in #3631
• feat: support namespace vended credentials for write by @jackye1995 in #5161
• docs: correct the comment in util.py by @shiyajuan123 in #5252
• feat: introduce community governance by @jackye1995 in #5262
• feat: add describe_indices function by @westonpace in #5221
• feat: add target_bases extension to python write_fragments API by @jaystarshot in #5234
• feat: support credentials vending for file reader and session by @jackye1995 in #5256
• feat: introduce blob arrow extension type by @Xuanwo in #5239

Bug Fixes 🐛

• fix: ensure recheck for IsNotNull in bloom filter by @Xuanwo in #5192
• fix: contributing URL gives 404 by @prrao87 in #5196
• fix: merge struct array use wrong child values by @wojiaodoubao in #5106
• fix: avoid unnecessary get_fragments calling during plan compaction by @zhangyue19921010 in #5179
• fix: split partition may be assigned to itself by @BubbleCal in #5190
• fix: improve schema validation for nullability and subschemas by @fenfeng9 in #4994
• fix: compile error in test_inline_transaction by @wojiaodoubao in #5206
• docs: fix batch udf with checkpoint document error by @fangyinc in #5185
• fix: memory-limited string BTREE index building by @wjones127 in #5175
• fix: home page code snippets cuasing various problems by @jackye1995 in #5245
• fix: panic if only one partition and split is triggered by @BubbleCal in #5241
• fix: clearer error in dataset take by @wkalt in #5243
• fix: docs and comment have broken links by @prrao87 in #5261
• fix: handle logical rows deletion properly for zonemap and bloomfilter by @HaochengLIU in #5140

Documentation 📚

• docs: introduce lance as a lakehouse format by @jackye1995 in #5209
• docs: update readme with latest lakehouse format info by @jackye1995 in #5216
• docs: capitalization change by @timsaucer in #5269

Performance Improvements 🚀

• perf: speed up filtered scan by up to 18.9× by moving the heavy CPU task out by @Xuanwo in #5165

Other Changes

• refactor: move blob version as a table level config by @Xuanwo in #5220

New Contributors

• @prrao87 made their first contribution in #5196
• @fenfeng9 made their first contribution in #4994
• @fangyinc made their first contribution in #5185
• @shiyajuan123 made their first contribution in #5252

Full Changelog: release-root/1.0.0-beta.N...v1.0.0-beta.4