You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@druid.apache.org by GitBox <gi...@apache.org> on 2019/01/23 01:19:31 UTC

[GitHub] peferron commented on issue #6889: [Proposal] Use bitmap iteration for filtered aggregators

peferron commented on issue #6889: [Proposal] Use bitmap iteration for filtered aggregators
URL: https://github.com/apache/incubator-druid/issues/6889#issuecomment-456629007
 
 
   I need to re-evaluate this in the context of #6794.
   
   Solution B yields good improvements with the current one-row-at-a-time query engine,but will likely not work as well with the vectorized query engine, since even low-selectivity filters are likely to have at least one matching row per batch (the default batch size in #6794 is 512 rows).
   
   On the other hand, it becomes possible to implement solution A on a per-batch basis. Right now, it looks like #6794 does not use bitmap iteration to build the list of matching rows yet, but checks the value of each row instead (see `SingleValueStringVectorValueMatcher`), so there might be room for improving filtered aggregator performance there as well. I'm going to try implementing that on top of the `vectoreyes` (what a pun @gianm) branch of #6794 to see how it goes.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@druid.apache.org
For additional commands, e-mail: commits-help@druid.apache.org