AlexanderSaydakov opened a new issue #12261: URL: https://github.com/apache/druid/issues/12261
We have noticed this recent change https://github.com/apache/druid/commit/e648b01afba52ee5b980d28b16b636e346a86819 This change introduced a piece of code which bypasses public API of DataSketches and accesses a private part of implementation. https://github.com/apache/druid/blob/033989eb1d8f4f91268b2d7d4d3dc73af7bf2c3f/extensions-core/datasketches/src/main/java/org/apache/druid/query/aggregation/datasketches/theta/SketchAggregator.java#L60 We believe that it is problematic and can lead to wrong results. We have this in our documentation: /** * Although the gadget object is initially an UpdateSketch, in the context of a Union it is used * as a specialized buffer that happens to leverage much of the machinery of an UpdateSketch. * However, in this context some of the key invariants of the sketch algorithm are intentionally * violated as an optimization. As a result this object can not be considered as an UpdateSketch * and should never be exported as an UpdateSketch. It’s internal state is not necessarily * finalized and may contain garbage. Also its internal concept of “nominal entries” or “k” can * be meaningless. It is private for very good reasons. */ -- 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]
