Short Updates
Thoughts on DuckDB's DuckLake so far:
1. As someone's who is familiar with the backend implementations of Delta Lake and Iceberg, the simplicity of DuckLake is admirable. 2. Iceberg table...
- As someone’s who is familiar with the backend implementations of Delta Lake and Iceberg, the simplicity of DuckLake is admirable.
- Iceberg table import/export is coming. Compaction support seems to be in the extension.
- I would love to see integrations for DuckLake with AWS Athena (Trino/Presto) and Spark. Implementors, just vendor the DuckDB implementation and call into it via JNI.
- The use of table level statistics is fantastic, join planning will perform well.
- Adding support for data sketches (bloom filters) on columns seems possible too!
Originally posted on LinkedIn.