Jackie-Jiang opened a new issue, #11706: URL: https://github.com/apache/pinot/issues/11706
When a group-by query does not have order-by on the aggregate column, we don't need to keep more groups than the LIMIT because the order-by value won't change. We can maintain a heap (PriorityQueue) of LIMIT values. On the group-key generation side, we should also keep only the relevant keys. One common query is: `SELECT COUNT(*) FROM myTable GROUP BY timeCol ORDER BY timeCol DESC LIMIT 10` Problems to solve: 1. Group-by query with order-by on the key column: Currently we keep `Math.max(5000, LIMIT * 5)` groups, which is not necessary since only the top `LIMIT` groups are relevant 2. Group-by query without order-by: Currently we keep random `LIMIT` groups per server, and there is no guarantee the same group is picked across different servers, which can lead to wrong result when there are more than `LIMIT` groups Solution: 1. To ensure the ordering is deterministic (we need this guarantee to ensure the groups returned from all servers are the same), we should append all non-ordering group keys implicitly. This is one exception: when we want to keep all groups on the server, we don't need this since all groups will be returned anyway. 2. Optimize the execution when all the ordering keys are group key -- 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: commits-unsubscr...@pinot.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
