jpountz commented on issue #12448: URL: https://github.com/apache/lucene/issues/12448#issuecomment-1643505863
Lazily heapifying sounds interesting, and thanks for sharing performance numbers when data occurs in random order. Do you also have performance numbers for the case when the index sort is the opposite order compared to the query sort? I'm curious how much this optimization can save in that case since this is what you're trying to optimize. > We dont have benchmark for numeric sort in Lucene itself Did you look at this task on nightly benchmarks? http://people.apache.org/~mikemccand/lucenebench/TermDTSort.html You might also be interested in checking out this [paper](https://www.vldb.org/pvldb/vol15/p3472-yu.pdf) where Tencent describes optimizations that they made for a similar problem in section 4.5.2: they configure an index sort by ascending timestamp on their data, but still want to be able to perform both queries by ascending timestamp and descending timestamp. To handle the case when the index sort and the query sort are opposite, they query on exponentially growing windows of documents that match the end of the doc ID space. -- 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
