walterddr commented on PR #11941:
URL: https://github.com/apache/pinot/pull/11941#issuecomment-1811799919
i think we can separate the 2 issues, IIUC the project ordering issue is
independent from the numSegmentsForThisPlan
1. the ordering issue will occur even if only 1 server is servering, thus
making sure the PROJECT node compare method in test is order invariant is the
main thing.
2. the numSegmentsForThisPlan IMO is really a "judgement call", b/c
- technically speaking we should return ALL plans if there are multiple
returns from different servers.
- however pinot doesn't really have a "logical" vs. "physical" plan
concept, where "logical" is invariant of the data layout; and "physical" should
be reported on every execution path.
- the result we are comparing here is neither truly logical or physical,
that it only return a single plan ("as logical") but it consolidates from
multiple server return ("physical").
hence my suggestion is we address the PROJECT ordering issue first in this
PR and address the plan issue separately
CC @somandal who improved upon this part of the logic to chime in :-)
--
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
