kishoreg commented on issue #7213: URL: https://github.com/apache/pinot/issues/7213#issuecomment-1012483684
@amrishlal Can you please address the comments from @Jackie-Jiang in the design doc? This is a complex feature and the design doc is mostly describing how Pinot operators work. It will help me a lot if you can expand on the following topics - How will the Window Operator work - Current proposed solution of pulling all records into a broker is not a scalable solution. Pulling all records into a single broker memory is not going to scale. - Without a distributed solution, the response times are going to be orders of magnitude slower than running it via Presto/Trino connector - In a typical database, Window operators are built on solid distributed partitioning, sorting primitives, and have some commonalities with primitives needed JOIN. Adding Joins first and then the window function is better in terms of sequencing. - Nothing related to the design but the estimates in the design doc gives me a feeling that we are grossly underestimating the complexity here Overall I am +1 on adding this feature but -1 on the design. -- 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]
