capistrant commented on issue #10042: URL: https://github.com/apache/druid/issues/10042#issuecomment-676836881
@leerho yes, once I did more examination, I found that I had not been aware of all of the druid query fragments sent off to data nodes. There were upgraded nodes returning hashes with no seedHash due to being empty sketches. For my organization, our sketches are either computed at query time on top of vanilla string dimensions or computed at ingestion time. We never have users ingesting pre-built sketches. Therefore, would I be correct in assuming that sketches built within druid at ingestion and query time would all use the default seed? If this is indeed, the case, I think my patch meets my specific requirements to bridge the gap to druid 18. As for if we want to commit this to the code base as something for all to use, I would be willing to work on that too. Are you suggesting something along the lines of the aggregator in the query spec specifying a seed that would be used in the case where we run into a sketch whose byte array is less than 16 bytes? ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
