Performance Improvements 🚀

  • Introduced two new settings, paradedb.topn_retry_scale_factor and paradedb.max_topn_chunk_size, which can be used to reduce the number of re-queries during a TopN when the index has lots of dead entries

Stability Improvements 💪

  • We now allow text key fields to use the keyword tokenizer
  • Fixed an index corruption issue related to the free space map
The full changelog is available here.