dsmiley removed a comment on issue #1270: LUCENE-9237: Faster UniformSplit IntersectBlockReader. URL: https://github.com/apache/lucene-solr/pull/1270#issuecomment-590966953 Your explanation really helps; perhaps it should be posted on the JIRA as it really summarizes what's going on. > For the proposal of the block prefix in the BlockHeader, does that mean that we have to open the block to get the prefix? Because the speed for FuzzyQuery highly depends on how many block we don't open. It would mean opening the block, admittedly, though it would allow short-circuiting reading the 4 blocklines before your code here gives up on the block. It's a shame we can't get better block common prefixes that allow short-circuiting fuzzy queries, at least an edit distance 1. I think it could help if the blockline chunking algorithm could consider the prefix in its algorithm so as to increase the probability of getting better block prefixes. It'd never be a guarantee since we want reasonable sized blocks.
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
