GoKawiilDuckDB 2.0 adds VARIANT type that 'shreds' JSON for faster queries
DuckDB 2.0 introduces VARIANT as a native data type, replacing text-based JSON storage. The engine analyzes JSON fields during checkpointing and extracts consistent fields—like a field that is always a number or always text—into real typed columns, leaving only the inconsistent or rare fields in a binary remainder.
GoKawiil's interpretation of the reporting above, not reported fact.
By avoiding full text parsing of JSON at query time, DuckDB could substantially speed up filters and aggregations on semi-structured data, a common bottleneck in analytics pipelines. This approach suggests DuckDB is positioning itself to compete more directly with systems that already offer native semi-structured storage, such as Snowflake's VARIANT type or BigQuery's JSON handling.
- VARIANT is now a first-class type in DuckDB 2.0, not just a JSON string column.
- DuckDB automatically 'shreds' consistent JSON fields into typed columns at checkpoint time.
- Only rare or type-inconsistent fields remain stored as binary blobs, reducing parsing overhead during queries.
Source: motherduck.com, 2026-10-10
Published there as: “Why DuckDB 2.0 is faster”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.