Marcel Kornacker has posted comments on this change.

Change subject: IMPALA-2805: Order conjuncts based on selectivity and cost
......................................................................


Patch Set 13:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/2598/13/fe/src/main/java/com/cloudera/impala/planner/PlanNode.java
File fe/src/main/java/com/cloudera/impala/planner/PlanNode.java:

Line 657:       double backoff_exp = 1.0 / (double) (sortedConjuncts.size() + 
1);
> I follow the intuition, but I think we'd need a more complex algorithm to t
the backoff affects the additive term (totalcost - e.cost), which is scaled by 
selectivity ^ backoff - it will decrease the impact of the selectivity of e, 
and that selectivity is different for each e. in other words, it will affect 
the cost calculation for each e differently.

i don't see how it would be immaterial to the ordering. that would only be the 
case if it added or subtracted a constant amount for each e.


-- 
To view, visit http://gerrit.cloudera.org:8080/2598
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-MessageType: comment
Gerrit-Change-Id: I02279a26fbc6308ac5eb819d78345fc010469034
Gerrit-PatchSet: 13
Gerrit-Project: Impala
Gerrit-Branch: cdh5-trunk
Gerrit-Owner: Thomas Tauber-Marshall <[email protected]>
Gerrit-Reviewer: Alex Behm <[email protected]>
Gerrit-Reviewer: Henry Robinson <[email protected]>
Gerrit-Reviewer: Marcel Kornacker <[email protected]>
Gerrit-Reviewer: Matthew Jacobs <[email protected]>
Gerrit-Reviewer: Mostafa Mokhtar <[email protected]>
Gerrit-Reviewer: Thomas Tauber-Marshall <[email protected]>
Gerrit-HasComments: Yes

Reply via email to