Joe McDonnell has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/15961 )

Change subject: IMPALA-9692 (part 3): Model QuerySchedule as a protobuf
......................................................................


Patch Set 4:

(1 comment)

http://gerrit.cloudera.org:8080/#/c/15961/4/be/src/runtime/exec-params.h
File be/src/runtime/exec-params.h:

http://gerrit.cloudera.org:8080/#/c/15961/4/be/src/runtime/exec-params.h@73
PS4, Line 73:   std::unordered_map<int32_t, const TPlanFragment&> fragments_;
> I guess the reason I did it this way is that the fragments don't come to us
What I had in mind is if you know the total number of indices, then you can 
presize the vector and assign at the indices as you go. That would require an 
initial pass to calculate the total size.

I don't have a strong preference. I think the vector is a bit more intuitive.

If you stick with the map, can you add a comment here saying that after 
initialization this map contains elements with keys 0 through size-1?



--
To view, visit http://gerrit.cloudera.org:8080/15961
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: I1db64e72f84604b1d8ac24e0bdd4ad6bedd6bcd9
Gerrit-Change-Number: 15961
Gerrit-PatchSet: 4
Gerrit-Owner: Thomas Tauber-Marshall <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Joe McDonnell <[email protected]>
Gerrit-Reviewer: Sahil Takiar <[email protected]>
Gerrit-Reviewer: Thomas Tauber-Marshall <[email protected]>
Gerrit-Comment-Date: Wed, 24 Jun 2020 19:58:39 +0000
Gerrit-HasComments: Yes

Reply via email to