tisonkun opened a new issue, #320: URL: https://github.com/apache/datasketches-rust/issues/320
## Problem A Theta union currently produces bytes through `union.to_sketch(ordered).serialize()`. Result extraction collects retained entries into an owned vector, applies the final theta and nominal-capacity rules, and optionally sorts it before the compact sketch writes another buffer. Callers that only need serialized results should not have to construct an independent compact sketch first. ## Suggested approach - Add direct union-result serialization with explicit ordered/unordered output semantics. - Share result preparation and encoding with compact-sketch serialization. Retain only the scratch storage needed for selection or sorting rather than requiring an owned intermediate sketch. - Preserve final-theta filtering, trimming to nominal capacity, canonical empty/singleton handling, and seed metadata. Serializing the internal hash table as-is is not correct. - Start with uncompressed union results. Also evaluate how the same internal encoding support can serve `ThetaIntersection` and `ThetaSketch::compact(...).serialize()`; these can be follow-up contributions. Intersection has an uninitialized result distinct from an empty sketch, which must remain explicit. Start with `datasketches/src/thetafamily/theta/union.rs`, `theta/sketch.rs`, and `common/union.rs`. PR #317 provides the HLL motivation. Keep any shared implementation internal; a new cross-sketch public serialization trait is not required. ## Validation Use the existing compact-result path as the reference. Cover ordered/unordered output, exact and estimation modes, empty and singleton results, mixed input precisions, and unions whose internal retained count exceeds nominal capacity. Verify compatibility fixtures and benchmark elapsed time and allocations. Establish any claimed performance improvement with measurements. Follow CONTRIBUTING.md for changelog entries, integration-test placement, serialization fixtures, and the required checks. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
