All pull requests

Pull requests

• feat(mem_wal): add a streaming to_reader to the MemWAL LsmScanner

  A-pythonenhancementK-approved
  #9694·zhangyue19921010 opened Oct 2, 2026Collaborator·

  ·

• fix(mem_wal): return JSON columns as Arrow JSON from LsmScanner

  bugK-approved
  #9693·zhangyue19921010 opened Oct 2, 2026Collaborator·

  ·

• fix: prevent incomplete unordered listings after retry errors

  A-encodingbugK-approved
  #9691·kamronis opened Oct 2, 2026·

  ·

• perf(knn): bound batched ANN candidate merge across segments

  K-approvedperformance
  #9688·kamronis opened Oct 2, 2026·

  ·

• fix(mem_wal): read frozen memtables in ShardWriter.lsm_scanner

  A-pythonbugK-approved
  #9687·zhangyue19921010 opened Oct 2, 2026Collaborator·

  ·

• fix(java): include baseId in DataFile/DeletionFile equals and add hashCode

  A-javabugK-approved
  #9686·Zhuoxi2000 opened Oct 1, 2026·

  ·
  1

  5

  

• fix(fts): preserve query position gaps in MemWAL phrase search

  bugK-approved
  #9685·lance-gatefixer[bot] opened Oct 1, 2026Bot·

  ·
  1

• fix(fts): reject incompatible segments before commit

  bugK-approved
  #9684·lance-gatefixer[bot] opened Oct 1, 2026Bot·

  ·
  1

• fix(fts): honor same-position alternatives in flat phrases

  A-indexbugK-approved
  #9683·lance-gatefixer[bot] opened Oct 1, 2026Bot·

  ·
  1

• fix(index): silence fieldless system index errors

  bugK-approved
  #9682·lance-gatefixer[bot] opened Oct 1, 2026Bot·

  ·
  1

• fix(dataset): refresh supplied manifests before commits

  bug
  #9654·geruh opened Oct 1, 2026Contributor·

  ·

• perf(scan): reuse fragment lookup when decoding plans

  performance
  #9653·geruh opened Oct 1, 2026Contributor·

  ·

• fix(scan): avoid exact counts when retaining deleted rows

  bug
  #9652·geruh opened Oct 1, 2026Contributor·

  ·

• fix: reject metadata ID allocation overflow

  bug
  #9651·geruh opened Oct 1, 2026Contributor·

  ·

• fix(table): validate fragment rewrite sources

  bug
  #9650·geruh opened Oct 1, 2026Contributor·

  ·

• perf(dataset): compute max field ID once per column update

  performance
  #9649·geruh opened Oct 1, 2026Contributor·

  ·

• feat(java): count indexed rows on explicit scalar segments

  A-javaenhancementK-approved
  #9644·hfutatzhanghb opened Oct 1, 2026Contributor·1 of 4 tasks·

  ·
  17

• test(encoding): concatenate struct test data with arrow's concat again

  A-encodingchoreK-approved
  #9643·slachiewicz opened Oct 1, 2026Contributor·

  ·

• perf(mem_wal): build LSM source scans with bounded concurrency

  K-approvedperformance
  #9642·beinan opened Oct 1, 2026Contributor·

  ·
  4

• fix(datafusion): reject a zero batch_size in LanceExecutionOptions

  bugK-approved
  #9641·slachiewicz opened Oct 1, 2026Contributor·

  ·

• refactor: stop using DataFusion APIs deprecated or changed in DataFusion 55

  K-approved
  #9640·slachiewicz opened Oct 1, 2026Contributor·

  ·

• fix(scanner): respect sliced list query offsets in vector search

  bugK-approved
  #9639·kamronis opened Oct 1, 2026·

  ·

• fix(index): apply segment ownership in shared-scan batch IVF search

  bugK-approved
  #9638·sezruby opened Oct 1, 2026Contributor·

  ·
  7

• fix(python): make maybe_sample return the requested number of rows

  A-pythonbugK-approved
  #9637·MohammadHijjawi97 opened Oct 1, 2026·

  ·

• fix(python): honor set_epoch in sharded samplers

  A-pythonbugK-approved
  #9636·MohammadHijjawi97 opened Oct 1, 2026·

  ·