somandal commented on PR #15817: URL: https://github.com/apache/pinot/pull/15817#issuecomment-2914332159
> * When there is single partition (`numPartitions == 1`), return 0 > * For RT segment following LLC segment name, use the one from segment name > * When `partitionColumn` is not null, try to find it from partition metadata, and throw exception if it cannot be found > * When `partitionColumn` is null, use the partition metadata if it only contains a single partition column > * Otherwise, fall back to the default hash One issue today is that for CONSUMING segments and REALTIME tables where there is not COMPLETED instance partitions (segment relocation disabled), we calculate partitionId based on segment last 4 steps above, even with numPartitions = 1. I think that's the right thing to do for this scenario. Whereas for COMPLETED segments (ONLINE and with segment relocation enabled), we set to partitionId = 0 if numPartitions = 0 to be consistent with OFFLINE tables. Similarly, StrictRealtimeSegmentAssignment mandates partitionId to exist and uses a different util to get the partitionId than other REALTIME tables So it is a bit harder to consolidate everything into a single utility without also passing some internal context -- 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 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
