gianm commented on PR #16073: URL: https://github.com/apache/druid/pull/16073#issuecomment-1985055000
> What do you think about some custom code into the MSQ layer that prevents resorting to the fields, in case the previous fields are already sorted by a superset of the given fields? For eg, if a stage wants to sort by columns A, B; while the previous single stage had sorted by columns A, B, C; then we can probably skip sorting for that stage. I think it makes sense as part of QueryKit. I don't think the `shuffleSpec` in the final `QueryDefinition`, once created by QueryKit, should be second-guessed by the MSQ controller. It looks like that's the path you took in that branch you linked, so that makes sense to me. Although, I don't understand why `aggregate` needs to be added to the hashed shuffle spec. Do you have an opinion on this patch? -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
