martijnvg commented on PR #14459: URL: https://github.com/apache/lucene/pull/14459#issuecomment-2804123675
Thanks for taking a look @stefanvodita! I've added the CHANGES entry. > Curious if you can share more about what you're trying to achieve! We have [a specialized doc values codec](https://github.com/elastic/elasticsearch/tree/main/server/src/main/java/org/elasticsearch/index/codec/tsdb/es819) for append only use cases that relies on index sorting. In order to speed up merging, we make use of the metadata statistics of the doc value producer in order to avoid pulling doc values instances from the producer and iterating over it multiple times. When indexing sorting is active, then this is more expensive. The consumer of this doc value codec creates a custom producer instance in the `merge*(...)` methods, which the `add*Field(...)` methods use to determine whether metadata statistics need to be computed or not. However the logic that creates the merged doc value instances and ordinal map are identical in the `DocValuesConsumer` class. This PR will allow us the reuse that logic and I think anyone creating a custom doc value codec with custom merge logic can benefit from this as well. -- 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: issues-unsubscr...@lucene.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
